LYMPH NODE SURGERY IN MELANOMA PATIENTS

There has been a historical emphasis on radical surgery both for the primary melanoma and for possible metastases within regional lymph nodes. This is because, in the absence of effective systemic therapy options, surgery was viewed as the only potential curative treatment.

There are alternate hypotheses for the role of surgery in lymph node metastases in melanoma. Firstly, there is the 'orderly progression' of the metastatic process, initially through locoregional lymphatics to the regional lymph nodes, and subsequently to the second-echelon lymph nodes and bloodstream. Therefore, early interruption of this process would in some way prevent further progression to distant visceral metastases and thereby improve overall survival (OS). The alternate hypothesis is that the development of metastatic disease within the locoregional lymph node(s) is an 'indicator' of metastatic potential and correlates with the development of distant metastases. However, removal of disease within lymph nodes at an early stage will not prevent the development of distant metastatic disease, which is governed by a complex interplay of the biology of the primary tumour and the host's innate defences.^[1] These two hypotheses are not mutually exclusive within a population of patients or indeed within an individual patient.

The first trials that analysed elective lymph node dissection (ELND) vs. nodal observation failed to demonstrate any survival advantages, and, after the introduction of the sentinel node (SN) procedure,^[2–6] ELND was abandoned because of prohibitive morbidity. It has been suggested that in these trials of ELND, the potential therapeutic effect of ELND on patients with lymph node metastases was not evident, because the majority of patients within the trials in fact did not have any lymph node metastases, which diluted the patients in whom a treatment effect might have been seen. The potential advantage of SN biopsy (SNB) is that it would be able to identify the correct subgroup of patients with nodal metastases who would potentially benefit from an early surgical intervention.

One area in which there is little debate is that the SNB procedure provides additional prognostic information over that available from analysis of the primary tumour.^[6,7]Furthermore, analysis of tumour burden within the SN is an even more accurate way of determining prognosis within the heterogeneous group of SN-positive patients. Patients with SN metastases > 1 mm have a poor prognosis, which is similar to that in patients with macroscopically (palpable or image detected) metastases.^[8–10]Although many different ways have been proposed to assess SN tumour burden, the most reproducible technique uses the Rotterdam criteria, adopted by the European Organisation for Research and Treatment of Cancer (EORTC).^[11] Many of the pivotal adjuvant therapy trials have used this 1-mm SN tumour burden cut-off for risk stratification based the Rotterdam–EORTC criteria.

The Multicenter Selective Lymphadenectomy Trial (MSLT)-I randomized patients (60 : 40) to wide local excision + SN [followed by completion lymph node dissection (CLND) in case of SN-positive disease] vs. wide local excision only (and nodal observation), with therapeutic lymph node dissection (TLND) upon the development of clinically apparent nodal disease on follow-up. The study failed to demonstrate a survival benefit of SN vs. nodal observation, with 10-year survival rates of 81·4% vs. 78·3% (P = 0·18).^[12]

A similar argument was made for the lack of a survival advantage in this study to that made for the trials of ELND – that the majority of patients in the study did not have lymph node metastases and so could not have gained any benefit from an intervention in the lymph node basin. Indeed, attempts at directly comparing the outcome of patients with nodal metastases identified either by SNB or clinically in follow-up appeared to show an improved outcome for patients undergoing SNB and CLND compared with TLND. Such direct subgroup comparisons may not be valid as they compared nonrandomized subgroups and might have included clinically irrelevant micrometastatic disease. Furthermore, the comparison excluded a small but important group of patients with false negative SNs who developed nodal disease on follow-up and had a very poor prognosis.^[13,14] Hence the predominant view after this study was that SNB provided valuable prognostic information, but that any effect on OS was not proven.

The question as to whether there is a therapeutic benefit of undergoing CLND for those patients identified by SNB as having nodal disease has also been analysed in randomized trials. Recently, two prospective randomized trials that examined the value of CLND in terms of survival vs. sequential nodal observation by ultrasound have reported their results. The German DeCOG-SLT study by Leiter et al. reported 3-year OS rates of 81·2% vs. 81·7% (P = 0·87) in 240 vs. 233 patients.^[15] This remained true after a median 72 months of follow-up, as presented at ASCO 2018. The MSLT-II trial confirmed this in 824 patients with CLND vs. 931 under observation only, with 3-year OS rates of 86% in both groups (P = 0·42).^[16]

Both studies did show improved relapse-free survival of the lymph node basin after CLND, which is to be expected after a prophylactic lymph node clearance in one of the two randomized groups. Predictably, the toxicity associated with the lymph node clearance was much greater in the CLND arm, in which all patients underwent a clearance, compared with the observation arm, where only those patients who relapsed underwent surgery. (Chronic lymphoedema was seen in 24·1% of the patients undergoing CLND vs. 6·3% in the observation group of the MSLT-II; P < 0·001.)^[16] Both studies had an over-representation of patients with SN tumour burden > 1 mm, which was likely due to patient and physician preference in case of larger SN tumour burden. However, the subgroup analysis of the more than one-third of patients in MSLT-II with SN tumour burden > 1 mm favours observation rather than CLND, indicating the high risk of systemic relapse in this group of patients.^[16]

Staging and loss of regional control have been issues raised to support the continued practice of CLND for SN-positive disease. Although 15–20% of SN-positive cases will show more nodes positive on CLND, this does not translate into upstaging of the same amount of patients.^[17,18] Two retrospective studies looking at this both demonstrated that only 6% of cases are upstaged by CLND, which does not seem to warrant this continued practice.^[17,18] Loss of regional control was also raised as an argument for CLND. This is because patients would present with bulky, no longer resectable disease, which would prevent adequate surgery or lead to an increased use of radiotherapy and thereby increase morbidity such as lymphoedema. The MSLT-II study did not show that a regional relapse equated to the loss of regional control, as the use of radiotherapy (8·1% vs. 6·5%) was less frequent in the observation group than in the CLND group.^[16,19] This also did not translate into more lymphoedema, as previously stated (6·3% observation vs. 24·1% CLND).

Recent developments have shown great improvement in OS and relapse-free survival for patients receiving adjuvant therapies such as ipilimumab, nivolumab, pembrolizumab, dabrafenib and trametinib.^[20–23] These therapies will soon become standard-of-care practice for the management of patients with stage III melanoma. Importantly, this will lead to a more routine use of SN staging for patients with clinical stage IB–II melanoma compared with the past decades, in which the use of the SN has been variable. Even though SNB in itself does not improve survival, it does identify the correct patients who might benefit from adjuvant therapy and whose survival will improve as a consequence of that treatment. SNB is likely to remain an entry criterion for adjuvant trials for patients with stage II melanoma to confirm that they do indeed have stage II disease.

Finally, some exciting results have been demonstrated for neoadjuvant immunotherapy for macroscopic (palpable) disease, with response rates of up to 80% being reported.^[24] Until now, these patients have still undergone lymphadenectomy after their neoadjuvant immunotherapy, but perhaps this is not necessary for all patients. A randomized trial is addressing this issue, by randomizing between CLND after assessing a complete response, and observation. This might reduce the extent of surgery for macroscopically (palpable node) involved patients in the future.

In conclusion, SNB is likely to take a central place in defining adjuvant treatment both for patients with melanoma in standard practice for patients with stage III disease, and in future clinical trials of patients with stage II disease. However, in the absence of a survival benefit, routine immediate CLND should be abandoned for the management of SN-positive disease in favour of ultrasound surveillance (Table 1).