New Zealand has the highest age-standardised incidence of melanoma in the world,1 with 2,567 new cases diagnosed in 2016.2 The treatment of melanoma also consumes significant healthcare resources, with the annual healthcare cost estimated to be upwards of 5.7 million dollars.3

Nodal tumour burden is an important prognostic marker in patients with melanoma, and can direct adjuvant treatment.4 Traditionally, sentinel lymph node biopsy (SLNB) is recommended for patients with primary cutaneous melanoma with a Breslow thickness exceeding 1.0mm. In patients with thin melanomas (less than 1.0mm) with ulceration or high risk features, a discussion regarding the benefits of SLNB is recommended.5–7 Until recently, most patients in New Zealand with a positive sentinel node would be counselled to undergo immediate completion lymphadenectomy.

Two recent randomised trials have challenged this traditional approach to SLNB management. The MSLT-II and De-COG SLT trials randomised patients with a sentinel node positive melanoma to immediate completion lymphadenectomy; or intensive clinical follow-up and delayed dissection for clinically detected recurrence. Both trials showed no improvement in overall disease-free survival in the immediate completion lymphadenectomy group.8,9 This has already led the Melanoma Multidisciplinary group in New Zealand to recommend intensive nodal observation for certain patients with a positive SLNB. In the nodal observation arm of the MSLT-2 study, patients were monitored with clinical examination and nodal ultrasound every four months for the first two years and every six months during years 3 to 5 of follow-up.

The cost and resource implications of this change in management are not well understood, and there has been no study in Australasia assessing this. It is likely that intensive nodal observation will become standard practice and it is essential that clinicians and managers anticipate this. The aim of this study is to assess the cost and resource implications of introducing intensive nodal observation for patients with cutaneous melanoma and a positive SLNB in a provincial New Zealand hospital.

Methods

Consecutive patients with a diagnosis of primary cutaneous melanoma and melanoma in situ presenting to Northland District Health Board between 1 January 2012 and 31 December 2014 formed the primary cohort. Patients were identified by searching the Faster Cancer Treatment Registry and the NDHB databases for inpatient admissions, incoming primary care referrals and incoming referrals from other health boards. Cases of ocular, intra-oral and anal melanoma were excluded. Cases of recurrent melanoma, as indicated by histopathology, were also excluded.

Patient electronic records were accessed using the electronic record and results reporting system, CONCERTO. Information collected included patient demographics and tumour characteristics, operations undertaken, outpatient clinic attendances and ultrasound attendances.

In order to reduce the effect of individual costing anomalies within our small cohort, we estimated the median cost for each type of melanoma-associated intervention from 1,012 patient episodes. This included all surgical admissions, day case procedures, outpatient clinic visits and outpatient ultrasound scans in Northland DHB from 1 July 2007 until 27 September 2018 that were associated with the primary diagnosis of melanoma. CostPro software was used to calculate in-house patient level costs for each episode and New Zealand common costing standards were used throughout. We were then able to apply these median costs to our primary cohort and a theoretical cohort undergoing intensive nodal observation in keeping with the regime outlined in the MSLT-2 trial. All patients in the intensive nodal observation group were assumed to survive three-year follow up to give a maximum estimated cost.

The primary area of interest was the additional clinical and financial resource required to introduce intensive clinical follow-up of the nodal basin for patients with positive sentinel nodes in cutaneous melanoma. Data was analysed in Microsoft Excel and presented using standard techniques.

The study was performed as part of a service review of melanoma care within NDHB. The study was reviewed by the Health and Disability Ethics Committee who issued an ‘out of scope letter’, deeming formal ethical approval unnecessary.

Results

Basic demographics and clinical outcomes

Two hundred and ninety-seven new cases of primary cutaneous melanoma were identified among 294 patients, including 82 patients with melanoma in situ. Median age at diagnosis was 69.1 years (IQR 58.9–76.5). The median length of follow-up was 1,737 days (IQR 1408–2024). Basic demographic characteristics of our primary cohort are outlined in Table 1. All patients with a positive sentinel node went forward for completion lymphadenectomy. This included three axillary, one neck and two groin dissections. Of the six patients followed up after immediate lymphadenectomy, four deviated from routine follow-up due to complications. These included seroma (N=3), wound infection (N=1), haematoma (N=1) and damage to the spinal accessory nerve (N=1). Four patients died during follow-up after immediate completion lymphadenectomy.

Table 1: Demographics, tumour characteristics and clinical outcomes within our primary cohort.

There were 29 patients eligible for SLNB on the basis of Breslow thickness who didn’t undertake it within Northland DHB. Nine did not proceed due to age or existing comorbidities, one patient suffered anaphylaxis to the blue dye, four patients were confirmed as having SLNB in the private sector. Fifteen patients were lost to follow up and it is not known how many of these went on to have private surgery elsewhere.

Costs and resourcing

Costs of traditional management were calculated for patients (N=6) in our retrospective cohort with a positive sentinel node biopsy (Table 2).Median length of follow-up post-lymphadenectomy was 2.60 years (IQR 1.32–5.21); mean 3.04 +/- 0.99 years.

Table 2: A cost and resource comparison of traditional management versus intensive nodal surveillance over three years.

‍

These were compared to a theoretical cohort undergoing intensive nodal surveillance over three years as per MSLT-2 protocol. The overall cost of traditional management in our retrospective cohort was $42,883 (mean cost per patient $7,147). All six patients undergoing intensive nodal surveillance were assumed to survive at least three years, resulting in a maximum estimated cost of $31,802 (mean cost per patient $5,300). This includes the forty-eight ultrasound scans required to follow them up over three years. A 26.1% nodal-recurrence rate over three years was predicted in this cohort, in line with the observation arm of the MSLT-2 trial. To provide context, the overall estimated cost to Northland DHB of melanoma excisions (including melanoma in situ) and inpatient admissions for SLNB and lymphadenectomy was $513,393 over the study period (Appendix).

Discussion

This study shows that in sentinel node positive melanoma, intensive observation of the lymph node basin is likely to be associated with lower costs than traditional management. However, there is a shift in resource requirements with lower operative and inpatient costs, and higher outpatient and imaging requirements.

Although our population was much smaller than the MSLT-2 study, the gender and tumour site distribution were comparable. Our rate of positive SLNB was 10.2%, compared to 20.8% in the published literature.10 Unlike MSLT-2, we also included seven patients with thin (0.75–1.0mm) melanomas with ulceration or high mitotic rate, all of whom went on to have a negative SLNB. It is worth noting that 10 patients did not undergo SLNB on clinical grounds. We are unable to estimate the effect of these patients or those lost to follow-up on our rate of SLNB positivity. However, we do not believe that they would alter the primary comparator of cost per patient in each cohort.

The benefits of lymphadenectomy are balanced against the risk of operative complications, which is known to be high. This is reflected in our study where patients required seroma catheter insertion, antibiotic therapy, percutaneous drainage or a prolonged course of physiotherapy. A drawback of our methodology is that these outpatient costs were not fully accounted for and the true financial burden of completion lymphadenectomy may be even higher than presented. It also demonstrates the importance of performing a true intention to treat analysis.

The higher sensitivity of ultrasound over clinical palpation in the detection of nodal recurrence is well established.11,12 However, sonographic criteria vary widely among published studies and no single feature is sufficiently specific. Ultrasound guided procedures are increasingly performed by surgeons in the clinic setting. However, the degree of interpretative experience required for ultrasound surveillance can only be expected from senior sonographers or radiologists. As a result, radiology departments will need to anticipate the increased demand that surveillance will have on their staffing and expertise.

Our primary area of interest in this study was the extra clinical and financial resource required to implement this treatment paradigm. We predicted that an extra five outpatient appointments and eight ultrasound scans would be required on average per patient over a three-year period. While this would require reallocation of resources within our hospital, the extra cost would be more than offset by the savings made from stopping immediate completion lymphadenectomy. In reality, these costs are low when considered in the context of the overall melanoma budget. Our estimated expenditure on melanoma exceeded half a million dollars over three years and didn’t take into account the countless negative excisional biopsies performed by our health board. Financial constraints should therefore pose no barrier to following new evidence-based guidelines in this setting.

The authors appreciate the limitations of this study, in particular in trying to predict cost and resource requirements from retrospective data. It is however the first study considering the resource implications of this clinical change in practice in New Zealand. Our population is small, but likely to be representative of other provincial DHBs across New Zealand. We also note that four of our completion lymphadenectomy patients died during follow-up, highlighting the poor prognosis among this group.

In short we have demonstrated that the introduction of intensive nodal observation in patients with positive SLNB in melanoma is likely to result in decreased costs, but some management or reallocation of resources will be needed. As the burden of patients with a positive SLNB is low, it is likely that the introduction of this type of regime would be achievable for provincial hospitals in New Zealand.

Appendix