With the increasing use of sentinel lymph node biopsy (SNB), the subsequent identification of nodal micrometastases has become the most frequent presentation of American Joint Committee on Cancer (AJCC) stage III melanoma. Although complete lymph node dissection (CLND) has long been regarded as the standard therapy for histologically confirmed stage III melanoma, the appropriate extent of surgery to manage micrometastatic disease diagnosed by SNB has become controversial. After a positive SNB, there is no residual disease identified on standard histopathology in the CLND specimen in 80%-86% of cases.1, 2 These findings led to the development of the Multicenter Selective Lymphadenectomy Trial II study.3 However, there are several reports suggesting negligible prognostic significance for minimally involved sentinel nodes, and several authors are already recommending no CLND.4, 5 In this environment, some surgeons are skeptical about the role of CLND in patients with a positive SNB. National Comprehensive Cancer Network (NCCN) data indicate that in the United States, only about 50% of melanoma patients with a positive SNB had a CLND in 2004-2005. The same study found that only 69% of cases had ≥10 lymph nodes removed when CLND was performed.6 These figures are of concern to the majority of specialist melanoma treatment centers that support the NCCN guidelines. These guidelines recommend thorough CLND for any positive SNB unless the patient is on an approved clinical trial or has serious medical comorbidities.7
In this issue of Cancer, Xing et al present data indicating a positive impact on survival of more thorough surgery in stage III melanoma patients. By using information from the Surveillance, Epidemiology, and End Results (SEER) population-based cancer registry,8 they report that a lower lymph node (LN) ratio, ie, the number of positive nodes divided by the total number of nodes removed, is a strong predictor of disease specific survival (DSS) for patients with primary melanomas of the head, neck, and upper and lower limbs. LN ratio is a continuous variable, but the authors recommended cut-points for the neck, axilla, and groin regional lymph node fields. The cut-point LN ratios identified on multivariate analysis were .07, .13, and .18, respectively. To remain below this cut-point ratio requires the removal of at least 15, 8, and 6 lymph nodes per positive lymph node for the respective CLND.9
The use of LN ratio is not a new concept, and an International Nodal Ratio Working Group has existed for several years.10 As referenced in Xing et al, an association between lower LN ratio and improved survival has been shown for cancers of the colon, lung, stomach, bladder, pancreas, and esophagus. It has also been reported for melanoma, breast cancer, and gynecological malignancy.9-13 Other authors investigating this association have evaluated LN ratio in various ways. Rossi et al used ≥10%, 10%-25%, and >25% in their evaluation of melanoma CLND specimens and found a relationship between LN ratio and overall survival.13 Katz et al reported that if there were ≥20% of nodes involved, patients had a significantly worse locoregional recurrence rate in breast cancer than those with <20% of lymph nodes involved.14
What does this mean? Does use of the LN ratio make sense when explained in terms of the surgical procedure undertaken and anatomical constraints? Is it plausible that, say, Person A with 10 nodes dissected from the groin and 2 nodes positive (LN ratio = 0.2) will have a DSS similar to that of Person B who has 40 nodes dissected from the groin with 8 nodes positive (LN ratio = 0.2)? Is the amount of surgery required proportional to the extent of lymph node involvement? For minimally involved nodal basins, is less than a thorough CLND adequate? Clearly this is not a practical solution. The surgeon does not know exactly how many lymph nodes are involved at the start of a CLND. Therefore doing less than an accepted minimum extent CLND is illogical. CLNDs are standard operations and their extent should be dictated by anatomical boundaries. They should be performed using techniques that are well described in the literature and hopefully taught to all surgeons who undertake them.15
On the other hand, the LN ratio concept may, to some extent, hold true for justifying extending the amount of surgery. For example, substantial clinical disease in the inguinal area is one of the commonly promoted indications for performing an ilio-inguinal dissection instead of just an inguinal dissection. This is the case even if there is no clinical or radiological evidence of pelvic lymph node involvement.7 In reality, this policy, and indeed the concept of the LN ratio, indicates that the more disease that is present in a lymph node basin, the more occult microscopic disease is also likely to be present. Therefore, not removing adjacent, clinically normal regional lymph nodes is likely to be leaving this subclinical disease in situ to progress and possibly metastasize further in the future. In terms of LN ratio, the surgeon may be excising too few nodes to ascertain the true number of positive nodes. This would reduce the denominator, increase the LN ratio, and, hence, put the patient in a worse outcome group with respect to DSS outcome.
The data from Xing et al and other publications on LN ratio support the concept that inadequate surgical clearance has higher rates of under-staging and resultant under-treatment. This is particularly pertinent in a disease such as breast cancer where adjuvant therapy is of proven efficacy and the extent of nodal involvement influences the amount of adjuvant therapy given. However, even in melanoma, where there is no highly effective adjuvant therapy, inadequate surgery seems to have a deleterious impact on both locoregional recurrence and DSS. This was proposed many years ago by Balch et al, who documented worse locoregional recurrence and median survival outcomes in a subgroup of patients who, on review of their management, had inadequate surgery during the first few years of recruitment to a South Eastern Cancer Study Group immunotherapy trial. The findings resulted in the establishment of minimum lymph node retrieval standards for CLND for patients to be eligible for inclusion in the group's future clinical trials.16 However, this and subsequent standards, such as those specified in the Multicenter Selective Lymphadenectomy Trials, have been based on expert opinion rather than compelling evidence.3 A recently published study from the Sydney Melanoma Unit evaluated 2269 CLNDs done by 9 surgeons.17 Each surgeon's performance was assessed, and a combined unit standard was also determined. This evaluation resulted in quality assurance parameters for the 3 main lymph node dissection fields to be used at the Sydney Melanoma Unit for ongoing audit. Table 1 compares these standards with the existing standards mentioned above.
|Axillary Dissection||Groin Dissection*||Neck Dissection†|
|Recommended LN retrieval||10||5||20|
|MSLT II-recommended LN retrieval||15||6||30|
|SMU 90th centile||10||7||20|
|Median number of LN||21||13||36|
|Frequency ≤10 LN removed||10%||25%||1%|
Axillary dissection is the CLND that demonstrates the least variance. The extent of groin dissection, and thus lymph node yield, varies with the frequency with which a surgeon does ilio-inguinal dissection compared with inguinal dissection. The extent of neck dissection can also be varied without untoward effect depending on the burden of disease and the known pattern of lymphatic mapping from the primary tumor site.15 When assessing the axillary lymph node retrieval by individual Sydney Melanoma Unit surgeons who had performed over 20 cases, the median lymph node retrieval varied from 19-27 lymph nodes. The overall median was 21 lymph nodes with an inter-quartile range (IQR) of 15-27.17 Xing et al reported the median axillary CLND lymph node retrieval was 15 with an IQR of 9-22.9 Kretchner et al reported that when lymphadenectomy was standardized in their unit, the median lymph node retrieval for axillary dissection increased from 6 to 12 lymph nodes.18 These variations in lymph node count can relate to the thoroughness of pathological assessment and perhaps the biological variation between the different population mixes in each study but are most likely related dominantly to variations in quality and extent of surgery. This is more likely to vary widely in a population-based database but should be more tightly monitored within a single center. If done as a prescribed operation with the recommended anatomical limits dissected, there seems to be a reasonably constant median and range of lymph node retrieval for each CLND. If a surgeon does enough cases to allow for this biological variation using a standardized technique and pathological assessment is made using standard protocols, then the yield should approach this standard lymph node number and range. If this occurs, then the relevance of the LN ratio may be lost because the denominator is constant and the variation in survival is again related only to the AJCC nodal stage of the tumor. This was our experience when LN ratio was evaluated using Sydney Melanoma Unit data (unpublished data). Others have theorized that the number of lymph nodes depends on the effectiveness of the immune response and that this relates to improved survival with lower LN ratio.19 This is possible but less likely and can be tested if surgery is standardized.
The study by Xing et al involved a retrospective assessment of the population- based SEER database and the largest sample size used to date to assess the impact of LN ratio and melanoma outcomes. An effort was made to adjust for differences in data coding across time points and for different regions. The statistical methods used were complex but cleverly applied. There are data available to adjust for a range of confounding factors. However, pooled data assessments always depend on the reliability of data for different surgeons using different techniques and over time. For example, the SEER database does not classify CLND by the operative field. The following assumptions are made: if the tumor was on the head and neck, then it drained to neck nodes; if it was on the upper limb, then it drained to the axilla; and if it was on the leg, then it drained to the groin. This led to the exclusion of all trunk melanomas from the study. It is well known that lymphatic drainage patterns vary and are sometimes unpredictable, and, therefore, these assumptions are bound to have an error rate.20
What is clear is that the current management of high-risk and resected metastatic melanoma is unsatisfactory because no highly effective adjuvant therapy is available. Therefore, surgery is often the only chance to cure AJCC stage III melanoma. Patients and surgeons wishing to avoid surgical treatment should understand this. Doing less than standard surgery should be reserved for appropriately powered, ethically approved clinical trials. An experienced team should complete surgery thoroughly and skillfully to maximize the potential for benefit. The data presented by Xing et al, and in other publications on LN ratio in melanoma as well as other malignancies, support the premise that thorough surgery is effective in improving locoregional control of disease and probably leads to improved survival outcomes.