Regional Lymph Node Surgery in Melanoma Patients

History and rationale

Surgical strategies for stage I and II patients have included two main components: wide excision of the primary tumor or biopsy site and regional lymph node evaluation. While recommendations for the extent of excision margins are well established and widely accepted, the approach to the clinically uninvolved regional lymph nodes has been the center of ongoing controversy. How to best manage the following clinical scenario is often called into question: A 36-year-old patient with a biopsy-proven 1.8 mm melanoma. Physical examination reveals the absence of enlarged lymph nodes in any potential regional lymph node group. The chest X-ray is normal and the patient is otherwise healthy. Traditionally, this patient would have been offered one of two options in addition to excision of the primary tumor: 1) observation of the regional lymph nodes and formal node dissection only if the patient subsequently develops clinically evident (palpable) nodal disease, an approach termed therapeutic lymph node dissection (TLND); or, 2) a formal lymph node dissection as a component of the initial surgical treatment, referred to as elective lymph node dissection (ELND). Both of these approaches have theoretical and very real disadvantages.

A significant percentage of individuals, predicted by increasing primary tumor thickness, ulceration, or other unfavorable histologic features of the primary tumor, harbor clinically undetectable regional lymph node metastases, which in most patients will progress to palpable (macroscopic) nodal disease if left untreated. Once clinical nodal involvement develops, the ability to achieve long-term survival and durable regional disease control with a TLND may be compromised compared to surgical approaches targeted at treating microscopic nodal burden. The rate of distant metastatic disease and relapse in the treated nodal basin is at least 50% and 15–50%, respectively, following TLND.

The practice of ELND was popularized for the sole intent of reducing these high rates of disease recurrence. Proponents of ELND suggested that removal of microscopically involved lymph nodes would prevent the development of clinically apparent lymph node disease, which in turn could, in a significant percentage of patients, eliminate a potential source of distant failure. Furthermore, a dissection performed when the lymph node involvement is microscopic would more completely eradicate regional metastases and prevent recurrence in the treated basin and the potential sequelae of pain, skin ulceration, blood vessel and nerve involvement, and advanced lymphedema that can be associated with this pattern of failure. In the majority of patients, however, microscopic nodal disease is absent at diagnosis and therefore they cannot benefit from an ELND and are subjected to the cost and morbidity of an unnecessary operation. It is therefore not surprising that overall survival advantage with ELND was not observed in prospective randomized trials that compared the outcome of stage I and II patients receiving either ELND or nodal observation.

For these reasons, the routine practice of ELND has been appropriately challenged. A rational compromise emerged when the technique of lymphatic mapping and SLN biopsy was introduced as a minimally invasive method for determining if occult nodal metastases are present. Patients with proven occult nodal disease in the SLN could then undergo an early TLND, and those without disease could be safely observed, an approach popularized as selective lymphadenectomy . This approach has been extensively studied worldwide.

Scientific support for the sentinel node concept

Lymphatic mapping relies on the hypothesis that the dermal lymphatic drainage from cutaneous sites to the regional lymph node basin is an orderly and definable process and that these lymphatic drainage patterns should mimic the metastatic spread of melanoma cells in the lymphatics ( Fig. 50.1 ). In this way, the first lymph node(s) receiving lymphatic drainage (the sentinel nodes) are the most likely to contain metastatic disease. The successful identification, surgical removal, and careful histologic examination of these nodes should provide accurate nodal staging.

To test this hypothesis, clinical studies were performed using intradermal injections of blue lymphatic dyes (isosulfan blue or patent blue V) at the primary tumor site followed by the visual identification of the SLNs in the nodal basin. These studies established the following: 1) SLN identification rates, and 2) the accuracy of the SLN in determining the presence or absence of regional nodal metastases.

The first report published by Morton et al. in 1992 evaluated 237 patients and demonstrated an 82% SLN identification rate. Subsequent studies from the M. D. Anderson and Moffitt Cancer Centers and the Sydney Melanoma Unit reported similar findings. Accuracy assessment was accomplished through the use of synchronous ELND performed at the time of the SLN biopsy. These initial studies evaluated 402 patients with successful SLN localization, 86 of whom were found to have regional node metastases (81 patients with a positive SLN and 5 additional patients with disease only in a non-SLN). This low false-negative rate of 5% supported the SLN concept.

Additional evidence that regional node metastasis is an orderly and non-random event was provided from the M. D. Anderson Cancer Center reporting on the examination of 105 lymphadenectomy specimens in patients with at least one positive SLN. Investigators found that the SLN was the only node involved in 83 (79%) of the basins, with disease in additional nodes identified in 21% of the lymphadenectomy specimens. Presented in another way, 68% of all the SLNs removed from all patients with at least one involved SLN and only 1.8% of all non-SLNs were involved with metastatic disease.

Tremendous interest was generated from these initial studies, and many centers subsequently adopted the selective lymphadenectomy approach for newly diagnosed intermediate- and high-risk stage I and II melanoma patients. Improvements in SLN localization techniques, insights into the biologic relevance of the SLN (discussed below), and additional findings supporting the SLN concept emerged. In a report of nearly 250 SLN-negative patients followed for over 3 years, only 10 patients (4%) developed nodal failure within the previously mapped regional basin. Such failures represent a false-negative rate similar to the 5% determined by concomitant ELND. More careful histologic scrutiny of the negative SLNs from these same 10 patients revealed the presence of disease in 8. These data not only further supported the validity of the SLN concept, but also suggested that routine histologic examinations of SLNs may fail to detect clinically relevant disease.

Technical advances

Initial SLN identification rates of 80% to 85% using blue dye injections provided a promising beginning. The use of high-resolution cutaneous lymphoscintography and an intraoperative handheld gamma detection device to locate radiolabeled colloids that have accumulated in SLNs after being injected at the primary site have yielded higher SLN identification rates. The use of a gamma probe was first described by Krag et al., who reported a 95% SLN identification rate. Studies comparing combined modality techniques (radiocolloid plus blue dye) versus blue dye alone demonstrated a significant increase in SLN identification to 99% with the combined approach.

The intraoperative use of the gamma probe provides a method of detection that is independent of and more sensitive than visualization of blue-stained nodes and can locate SLNs that might otherwise be undetected. The number of SLNs identified was greater when both modalities were employed compared to blue dye alone (1.74 versus 1.31, respectively). This more complete removal of SLNs may further reduce the already low false-negative rate. Figure 50.2 summarizes the components necessary for successful identification and removal of an SLN.

These techniques can also aid in the localization of SLNs that may exist outside and/or proximal to the formal nodal basin; referred to as interval, in-transit, or ectopic SLNs ( Fig. 50.3 ). According to published studies, the frequency of such SLN locations is in the range of 5% to 10% of patients, and the frequency of involvement with microscopic disease is the same as that of SLNs harvested from formal basins. The failure to identify these nodes risks under-staging some patients and leaving behind potential sources of clinical recurrences.

SLN biologic and prognostic significance

Studies have demonstrated that the incidence of SLN metastases correlates directly with increasing tumor thickness ( Table 50.1 ). SLN involvement is also associated with a variety of other known primary tumor factors predictive of overall survival, including ulceration, lymphatic invasion, mitotic rate, Clark level, anatomic site, and host factors such as age. In a multivariate analysis, the two variables that independently best predicted SLN involvement were tumor thickness and the presence of ulceration. This analysis uncovered a unique interaction between tumor thickness and ulceration in that the presence of ulceration within a specific tumor thickness stage worsened the prognosis of patients equivalent to those in the next-higher thickness group without ulceration. A similar relationship between thickness and ulceration in terms of predicting the incidence of SLN metastases exists, as shown in Table 50.1. These observations support the hypothesis that the prognostic value of tumor thickness and ulceration is largely dependent on the fact that these two same factors predict SLN metastases, and in this way offers convincing evidence that SLN involvement is a biologically important event.

Further supporting this conclusion are findings from survival analyses of large numbers of stage I and II patients managed in prospective selective lymphadenectomy programs. Consistently, these reports revealed that the SLN-positive patients experienced a significantly lower survival compared to SLN-negative patients ( Fig. 50.4 ), and that the histologic status of the SLN was the most powerful independent predictor of overall survival in the clinically node-negative melanoma patients when analyses were carried out including previously described primary tumor prognostic factors ( Table 50.2 ). Several similar analyses have since been published from large single institutional as well as multi-centered experiences corroborating these findings.

While the patients with a negative SLN as a whole enjoy an excellent survival, a negative SLN is not a perfect prognostic factor. Five-year melanoma-specific survival rates are generally 90% for the SLN-negative patients, with recurrence and death occurring secondary to the following reasons: a false-negative SLN or pure hematogenous pattern of metastases. Predictors of relapse and death in the SLN-negative group include increasing tumor thickness and primary tumor ulceration. The development of clinical nodal disease in a nodal basin previously determined to be without microscopic involvement defines a false-negative event and occurs in approximately 3–5% of these patients. Theoretically, three explanations exist for these events: 1) the main SLN was not properly identified during the SLN procedure, leaving behind a microscopically involved lymph node; 2) the original SLN procedure was accurate, but microscopic in-transit disease was present at the onset that had not yet traveled to the nodal basin; 3) the correct SLN was removed and microscopic disease was present but undetected by the histologic examination either as a result of a very small burden of disease or within a portion of the node that was not sampled.

Rationale for SLN biopsy

The original motivation to study SLN biopsy was to establish an effective method of preventing the development of clinically palpable regional disease in the stage I and II melanoma patients. Many have questioned its therapeutic role. Results of recent clinical trials demonstrated survival benefits and improved regional disease control afforded by SLN dissection in patients with microscopic nodal involvement, and improved survival with the use of adjuvant interferon in node-positive patients. These studies provided additional motivation to incorporate SLN biopsy in the management of stage I and II patients.

Does early node dissection impart a survival benefit?

The potential for improved survival with early node dissection was the goal for the routine application of ELND as part of the initial management of newly diagnosed stage I and II patients. The question of survival impact with the use of ELND relative to nodal observation and therapeutic dissection for those patients who develop clinically detectable nodal disease has been evaluated in four prospective randomized trials. The first two trials, one from the Mayo Clinic and one from the World Health Organization, performed in the 1970s and prior to knowledge concerning primary tumor prognostic factors, demonstrated no survival advantage. Accordingly, ELND was strongly contested and largely abandoned. These trials were subsequently criticized because the study populations were at low risk for occult nodal disease and therefore unlikely to benefit from the surgical treatment being tested.

Two additional ELND trials were performed targeting the higher-risk clinically node-negative patients. Trends for improved survival following ELND were observed in both trials. However, these differences were not statistically significant. While many concluded that early treatment of nodal metastases had little impact on disease progression, others suggested that these trials were underpowered because only the 20% of patients harboring nodal disease could potentially benefit from the procedure. Long-term results published in 1998 from the WHO ELND Trial, which included patients with trunk primaries >1.5 mm, demonstrated that patients with microscopic nodal disease in the ELND treatment arm experienced improved overall survival compared to patients who developed clinical adenopathy after randomization to excision alone. Results published in 2000 from the Intergroup ELND Trial in which patients with melanomas 1 to 4 mm in thickness were studied, demonstrated that prospectively stratified subgroups (1 to 2 mm and all non-ulcerated primaries) derived a survival benefit with ELND.

While overall survival rates for the entire study cohorts in both trials were not statistically different (confirming that not all patients can benefit from ELND), these studies do suggest that specific subsets of patients (most notably those with microscopic nodal disease and possibly additional patients with nodal disease undetected by routine histologic techniques) can benefit from earlier dissections. These data offer evidence-based credence to the theoretical concerns of delaying the lymphadenectomy until palpable nodal disease develops and supports the selective lymphadenectomy approach.

The survival impact of the selective lymphadenectomy strategy, using SLN biopsy as an alternative to ELND, was formally studied in a prospective randomized multi-centered international trial comparing the outcomes of nodal observation after wide excision to SLN biopsy and completion dissection for patients with microscopic nodal involvement. The design and primary and secondary endpoints of the Multicenter Selective Lymphadenectomy Trial-1 (MSLT-1) are presented in Figure 50.5 . The results of the third interim analysis of the MSLT-1 were recently published in the New England Journal of Medicine. Data were available for 1269 patients. In the biopsy group, the presence of metastases in the SLN was the most important prognostic factor. The 5-year melanoma-specific survival rate was 72.3±4.6% among patients with tumor-positive SLNs and 90.2±1.3% among those with tumor-negative SLNs ( P <0.001), confirming the previously reported observations from several other groups. The melanoma-specific death rate at 5 years was similar in the two groups (13.8% in the observation group and 12.5% in the biopsy group), as was the melanoma-specific survival rate at 3 years (90.1±1.4% and 93.2±0.9%, respectively) and 5 years (86.6±1.6% and 87.1±1.3%, respectively). The incidence of SLN micrometastases was 16%, while the rate of relapse in regional nodes in the observation group was 15.6%. The mean number of tumor-involved nodes at lymphadenectomy was 1.4 in the biopsy group and 3.3 in the observation group ( P <0.001). A pronounced overall survival advantage was observed when the analysis was performed including only the node-positive patients. Compared to the patients who underwent a therapeutic (delayed) dissection for clinical nodal failure after being randomized to nodal observation, the SLN-positive patients who underwent immediate lymphadenectomy enjoyed a 20% improved 5-year survival rate (72.3±4.6% versus 52.4±5.9%; hazard rate for death 0.51, 95% CI 0.32–0.81; P =0.004) ( Fig. 50.6 ).

The interim results of the MSLT-I Trial provide important insights into the value of selective lymphadenectomy compared with delayed lymphadenectomy. The lack of an overall survival difference between the two treatment arms is not surprising in that this trial suffers from the same limitations as the ELND trials: being underpowered because of the low percentage of patients (16% in this trial) who could benefit from complete lymphadenectomy. Assuming that early lymphadenectomy for SLN-positive patients is associated with a 20% survival benefit, one would predict an overall survival advantage of no more than 3.2% compared with delayed lymphadenectomy. Nonetheless, survival differences can emerge with longer follow-up. If future events follow the patterns observed in the two ELND trials, more recurrences in the nodal observation arm may develop over time than in the SLN biopsy arm.

The results of the secondary survival analysis comparing SLN-positive patients with those who developed clinically palpable nodes following nodal observation are particularly noteworthy. The improved survival of the SLN-positive group not only corroborates the results of the WHO ELND trial but also supports the concept that, if left intact, microscopic nodal disease progresses and is associated with a worse prognosis. In some patients, therefore, increasing nodal burden can be a source of systemic dissemination; early treatment of nodal disease can favorably alter the natural history of their disease.

Regional disease control

The most common first recurrence in primary melanoma patients initially treated with excision of the primary site alone is palpable lymph node metastases. These patients are generally treated with a TLND for attempts at cure and regional control of disease. Reported in-basin, post-dissection failure rates range from 9% to 50% depending on a variety of factors, including basin site, number and size of involved nodes, and presence of extracapsular extension. In-basin recurrences are very difficult to treat surgically and may be the source of significant morbidity in the form of pain, severe lymphedema, venous obstruction, skin ulceration, nerve involvement, and bleeding. In-basin failures in patients treated with ELND and found to harbor microscopic disease occur in less than 10% of patients and are reported to be even lower after completion dissection in SLN-positive patients. The potential for improved regional disease control when dissections are performed for microscopic disease further supports the use of SLN biopsy.