Treatment patterns, outcomes, and resource utilization of patients with metastatic melanoma in the U.K.: the MELODY study


  • Funding sources This work was supported by Bristol-Myers Squibb. Representatives from Bristol-Myers Squibb were involved in the study design; in the collection, analysis and interpretation of data; in the writing of the manuscript; and in the decision to submit the manuscript for publication and have been listed as authors of this manuscript. Editorial and writing assistance was provided by StemScientific, funded by Bristol-Myers Squibb. None of the authors received financial compensation for authoring the manuscript.
  • Conflicts of interest P.L. has acted as a consultant/advisor to Bristol-Myers Squibb, Roche, Novartis, Celgene and GSK, and has also received travel support from Bristol-Myers Squibb and Roche. M.M. has received travel support from and acted as a consultant/advisor to Bristol-Myers Squibb. R.O. has acted as a consultant/advisor to Bristol-Myers Squibb. At the time of writing, U.B. and Q.W. were employees of Bristol-Myers Squibb. Q.W. is currently at GlaxoSmithKline, Stockley Park, Uxbridge, U.K. L. P. is currently at Novo Nordisk, Copenhagen, Denmark.



Advanced melanoma is an aggressive disease with a poor prognosis. Approved therapy is limited in the U.K. and, until recently, no treatment had improved survival over best supportive care. A deeper understanding of current clinical practice will help new agents find a place in future treatment pathways.


To document U.K. clinical practice for the treatment of patients with unresectable stage III/IV (advanced) melanoma.


MELODY (melanoma treatment patterns and outcomes among patients with unresectable stage III/IV disease: a retrospective longitudinal survey) compiled registries of consecutive patients with malignant melanoma (any stage) between 1 July 2005 and 30 June 2006 from France, Italy and the U.K. Patients with advanced melanoma and ≥ 2 months of follow-up were eligible for analysis.


There were 220 eligible patients identified in the U.K., of whom 117 (53·2%) received systemic therapy outside of clinical trials. Over half of these patients received dacarbazine as first- or second-line therapy. Healthcare-resource utilization was extensive and patients had short survival times: 1- and 2-year survival rates after first-line systemic treatment were 45·5% [95% confidence interval (CI) 37·1–53·6] and 24·7% (95% CI 17·7–32·3), respectively.


Systemic and palliative treatments used to manage advanced melanoma in the U.K. are associated with considerable healthcare resource utilization and poor short-term survival.

Malignant melanoma accounts for approximately 4% of all new cancer cases in the U.K. Each year, there are over 12 000 new cases of malignant melanoma and over 2000 deaths.[1] While the incidences of stomach, colon, lung, cervical and brain cancers are predicted to decrease (10–20%), the incidence of melanoma is predicted to increase (30–40%) by the year 2020.[2] This increasing incidence is accompanied by plateaus and decreases of mortality rates in some European countries, possibly related to earlier recognition of the disease.

Local recurrences of melanoma and in-transit metastases are most effectively treated by surgical excision, while surgery or radiotherapy are useful options for selected lesions in oligometastatic disease.[3] In general, patients with metastatic melanoma have a very poor prognosis, with median survival of 6–10 months after diagnosis of metastatic disease. However, this may improve as newly approved agents become used more widely.[4-9] Melanoma has a strong tendency to metastasize to the brain, whereupon the already poor prognosis worsens.[10, 11]

Until recently, U.K. patients with unresectable metastatic spread had very few systemic treatment options.[12] Despite no demonstrable survival benefit, chemotherapy is widely used for patients with unresectable stage III or stage IV (advanced) melanoma, and this remains the case for those not eligible for targeted therapy, with dacarbazine (approved) or temozolomide (off-label) being standard first-line agents.[13-15] Existing guidelines recommend that patients with stage IV disease are considered for entry into clinical trials.[12] Because of the paucity of approved and effective treatment options, clinicians have often adopted different approaches to disease management, including experimental regimens or variations to palliative or best supportive care.[16]

Recently, two agents with different mechanisms of action have improved survival in phase III studies of patients with advanced melanoma.[5, 7, 17] Ipilimumab, a fully human monoclonal antibody against cytotoxic T lymphocyte-associated antigen-4, was approved by the European Union in July 2011 for the treatment of advanced (unresectable or metastatic) melanoma in adults who have received prior therapy.[18] Vemurafenib, an inhibitor of BRAF (serine/threonine-protein kinase B-Raf), was approved for the treatment of adult patients with BRAF V600 mutation-positive unresectable or metastatic melanoma by the European Union in February 2012, alongside a test to identify the BRAF V600 mutation.[19]

Understanding the current disease burden and unmet medical need of patients with advanced melanoma provides context for defining the role of new agents such as ipilimumab and vemurafenib in the treatment pathway. Although the U.K. has comprehensive cancer registration systems in place, data on patients with advanced melanoma are not readily accessible. Cancer registries require diagnostic, tumour and treatment details to be collected, but only details of death are required in terms of outcomes data. Recently, the U.K. National Cancer Intelligence Network stated that more complete registration of skin cancers was necessary to support service delivery and inform the development of a national prevention initiative.[20]

MELODY [melanoma treatment patterns and outcomes among patients with unresectable stage III/IV disease: a retrospective longitudinal survey (CA184-068; Bristol-Myers Squibb)] was the first multinational (Italy, France and the U.K.), observational study to characterize treatment patterns and associated outcomes for patients diagnosed with advanced melanoma. The primary aim of the study was to document current treatment patterns in patients with advanced melanoma, particularly outside of clinical trials where current data are lacking. The study also captured information on patient and disease characteristics, clinical outcomes and healthcare resource utilization.

Pan-European data from MELODY have been published.[21, 22] These data confirmed that, despite variations in treatment patterns and healthcare resource utilization between the countries, there was an urgent need for more effective treatment options in patients with advanced melanoma. Here, a detailed U.K.-specific analysis is presented, including data on treatment patterns, outcomes and resource utilization, previously published in part as a congress abstract.[23] As far as possible, the principles outlined in the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) Statement[24] have been adhered to.

Patients and methods

Study design and data collection

U.K. data were collected as part of a multinational, observational, retrospective, longitudinal survey of patients diagnosed with unresectable stage III/IV melanoma. Site assessments were performed to collect site information on geographic location, type of hospital (general district, academic, specialized), number of patients with melanoma treated and level of interest in participating in the study, including the ability to retrieve and access medical records during the time of interest and the availability of site resources. Consecutive patients with a diagnosis of malignant melanoma (stage I–IV) who attended a participating site at least once between 1 July 2005 and 30 June 2006 were entered into a registry that captured a limited set of parameters related to date and stage of disease. Of these patients, those with a diagnosis of unresectable stage III/IV (advanced) melanoma who had ≥ 2 months of follow-up were eligible for analysis (Fig. 1). Data from patients with advanced melanoma and < 2 months of follow-up were captured in the screening log to assess potential bias against patients with aggressive disease or poorer prognosis.

Figure 1.

Patient numbers in selected categories. Screened patients refers to all patients with melanoma (any stage) who attended a participating site at least once between 1 July 2005 and 30 June 2006.

Medical charts were abstracted for eligible patients from the date of diagnosis of advanced melanoma until 1 May 2008 or death, whichever occurred first. Abstracted data were documented in a standardized case report form, completed by the relevant healthcare professional. The study was conducted in accordance with the principles of the Declaration of Helsinki, Good Clinical Practice, Good Epidemiological Practice and applicable regulatory requirements. Written informed consent was not required, as per local requirements for the collection of fully anonymized data for observational research. The study protocol was reviewed and approved by a duly constituted independent ethics committee before patients were screened for entry.


The primary objective was to document first-line treatment use. Secondary objectives were to estimate the proportion of patients with unresectable stage III/IV melanoma at first diagnosis; document patient and disease characteristics (age, sex, date and disease stage at first melanoma diagnosis, date and disease stage at advanced melanoma diagnosis, location of primary tumour, sites of metastatic disease, performance status); assess patient outcomes as documented by the treating clinician [response rate, disease control rate, time to response, duration of response, progression-free survival (PFS), overall survival (OS)]; and determine the proportion of patients who progressed through first-line treatment and describe second-line and subsequent treatments, and ascertain healthcare resource utilization [hospitalizations, hospice care, outpatient visits, emergency department visits, management of adverse events (AEs)] related to melanoma treatment. Only resource use associated with common AEs (transfusion, administration of concomitant medications including antiemetics and growth factors) outside of clinical trials or as part of supportive care was captured.

Statistical analysis

Analyses were predominantly descriptive. For categorical variables, counts and percentage of patients in each category were presented. Missing data were not included. Two main datasets were used for analyses: the overall melanoma population (i.e. total of site registry entries), and the eligible melanoma population, comprising patients with advanced melanoma with ≥ 2 months of recorded follow-up.

Overall survival and PFS were analysed using Kaplan–Meier methodology. OS was determined for all eligible patients from the date of advanced melanoma diagnosis to date of death or 1 May 2008, whichever occurred first, and for the population of patients who were first diagnosed with advanced melanoma between 1 July 2005 and 30 June 2006. OS was also calculated for both cohorts (excluding patients with brain metastasis or ocular melanoma) from the start date of first- and second-line systemic therapy until date of death or 1 May 2008.


Patient disposition

From a proposed list of approximately 55 sites, 10 sites in the U.K. participated in MELODY, collectively screening 1048 patients (Table 1). Of these, 277 patients (26·4%) had advanced melanoma during the inclusion period (Fig. 1) and 47 (4·5%) had advanced melanoma at first diagnosis. Of the 277 patients with advanced melanoma, 220 had ≥ 2 months' follow-up and were eligible for analysis; 57 (20·6%) patients were not included in follow-up. Patient and disease characteristics are summarized in Table 2.

Table 1. Number of screened and eligible patients for analysis at sites across the U.K
SitePatients screened (= 1048)Patients eligible for analysisa (= 220)
  1. a

    Presented with unresectable stage III/IV melanoma during the inclusion period and had ≥ 2 months' follow-up.

St Thomas' Hospital10926
St George's Hospital3225
Table 2. Patient and disease characteristics of 220 eligible cases
  1. ECOG PS, Eastern Cooperative Oncology Group Performance Status; TNM, tumour, nodes, metastasis. aFive patients were assessed using the Karnofsky performance status, which was 70–100% for all.

Mean (median) age at the point of inclusion, years58 (59)
Male/female, %56/44
Median (mean) time from first diagnosis to diagnosis of advanced melanoma, years2·1 (3·7)
Stage at first diagnosis, n/N (%)
Stage I, II, III (resectable or status unknown)158/188 (84)
Unresectable stage III/IV30/188 (16)
Missing32/220 (15)
Low/normal/elevated serum lactate dehydrogenase level at the point of inclusion, %1/67/33
ECOG PS prior to first-line therapy,a n/N (%)
0111/171 (65)
152/171 (30)
26/171 (4)
3–52/171 (1)
Missing44/220 (20)
TNM classification: metastatic staging, n/N (%)
M020/103 (19)
M1a11/103 (11)
M1b20/103 (19)
M1c42/103 (41)
Unknown10/103 (10)
Missing117/220 (53)
Location of primary tumour, n (%)
Extremities91 (41)
Trunk63 (29)
Head and neck17 (8)
Acral5 (2)
Mucosal1 (0·5)
Other/unknown40 (18)
Sites of metastatic disease, n (%)
Bone22 (10)
Lymph nodes106 (49)
Liver55 (26)
Lung103 (48)
Brain33 (15)
Other92 (43)
Not applicable1 (0·5)

Treatment patterns

Among the 220 eligible patients, 205 (93·2%) received first-line treatment, 80 (36·4%) continued to second-line treatment and 30 (13·6%) received third-line treatment. The most common treatment was systemic (Fig. 2); 145 patients (65·9%) received first-line, 67 (30·5%) received second-line and 26 (11·8%) received third-line systemic treatments. Of 144 patients who received systemic first-line treatment (data from one patient were missing and excluded from analyses), 47 (32·6%) received their treatment within a clinical trial. Respective numbers for patients receiving second- and third-line systemic therapy within a clinical trial were 27 of 67 (40%) and seven of 26 (26%). Across all lines of therapy, 117 of 220 patients (53·2%) received systemic therapy (chemotherapy, immunotherapy or combinations of both) outside a clinical trial, including 84 of 117 patients (71·8%) treated with dacarbazine. Across first- and second-line therapy, 29 of 220 patients (13·2%) received immunotherapy outside a clinical trial (Fig. 2). The most commonly used systemic treatments outside clinical trials are shown in Table 3.

Table 3. Commonly used systemic treatment options and response rates outside a clinical trial
  1. CR, complete response; PR, partial response; DCR, disease control rate. aN = 96 (two missing); bN = 39 (one missing); cfirst-line use reported mostly from one centre.

Patients receiving systemic treatment, n/N (%)98/220 (44·5)40/220 (18·2)
Median duration of systemic therapy (weeks)11·3a6·6b
Top three systemic treatments at each line of treatment, n (%)
Dacarbazine63 (64)21 (52)
Vindesine 5 (12)
Interferonc17 (17)4 (10)
Temozolomide7 (7) 
Carboplatin 4 (10)
CR/PR, n (%)22 (23)6 (15)
DCR, n (%)37 (38)11 (28)
N (missing)97 (1)40 (0)
Time to best response (months)
Median (min; max)4·8 (1·6; 19·1)2·2 (1·4; 2·9)
N (missing)22 (0)6 (0)
Duration of best response (months)
Median (min; max)4·4 (0; 21·5)1·4 (1·0; 3·3)
N (missing)17 (5)5 (1)
Figure 2.

Overview of patient treatment in 220 eligible U.K. cases. aOne missing for first-line treatment type data. Tx, treatment.

Survival and response rates

Among patients who received systemic treatment outside of a clinical trial at any line (= 117), median PFS was 2·8 months [95% confidence interval (CI) 2·3–3·9]. Data for OS from time of diagnosis of advanced melanoma to death/1 May 2008 were available for 216 patients. Median OS was 15·6 months (95% CI 12·6–17·8). As expected, survival was shorter among patients with elevated lactate dehydrogenase (LDH), lymphopenia, brain metastases and more advanced disease at baseline; however, the 95% CIs overlapped in each case.

Excluding patients with brain metastases and ocular melanoma, median OS from the start of first-line therapy (= 129) was 11·0 months (95% CI 8·8–13·5), with 1-year and 2-year survival rates of 46·2% and 26·5%, respectively. For the 59 patients receiving first- and second-line systemic therapy, median OS from the start of second-line therapy was 6·9 months (95% CI 5·3–9·8), with 1-year and 2-year survival rates of 28·1% and 13·8%, respectively. Analysis of patients diagnosed with advanced melanoma during the inclusion year (= 144) showed similar trends to that of all eligible patients; however, survival was lower across all lines of therapy (Fig. 3 and Table 4).

Table 4. Survival analysis
 Patients, nMedian OS, months (95% CI)1-year survival rate, % (95% CI)2-year survival rate, % (95% CI)
  1. OS, overall survival; CI, confidence interval. aExcluding patients with brain metastases and ocular melanoma; bBetween 1 July 2005 and 30 June 2006.

Among all eligible patients
From date of diagnosis21615·6 (12·6–17·8)58·1 (51·0–64·5)32·6 (26·2–39·1)
From first-line treatmenta12911·0 (8·8–13·5)46·2 (37·3–54·7)26·5 (19·0–34·7)
From second-line treatmenta596·9 (5·3–9·8)28·1 (17·0–40·3)13·8 (6·0–24·7)
Patients diagnosed with advanced melanoma during the study inclusion yearb
From date of diagnosis14410·1 (8·4–13·5)44·7 (36·2–52·9)22·4 (15·6–30·0)
From first-line treatmenta868·6 (6·6–11·0)35·7 (25·5–46·0)19·0 (11·2–28·3)
From second-line treatmenta325·6 (4·0–9·8)16·7 (6·1–31·9)8·4 (1·6–22·6)
Figure 3.

Kaplan–Meier analysis for overall survival (OS) among 220 eligible patients from date of diagnosis of advanced melanoma and 144 eligible U.K. patients diagnosed with advanced melanoma between 1 July 2005 and 30 June 2006. CI, confidence interval.

Clinical response data were available for 117 patients who received systemic treatment outside a clinical trial. A complete or partial response was reported in 22 of the 98 patients (23%) during first-line systemic treatment and six of the 40 patients (15·0%) during second-line systemic treatment. Corresponding disease control rates were 37 of 98 (38%) and 11 of 40 (28%), respectively. Responses were of short median duration (4·4 and 1·4 months, respectively) (Table 3).

Resource utilization

Healthcare-resource utilization was determined for patients who received systemic treatment outside a clinical trial or supportive care. Patients were more likely to receive hospice care or be hospitalized during supportive care than during active (systemic) therapy (Table 5). The median total durations of hospitalization and hospice care were 14 days [interquartile range (IQR): 6–23 days; = 42 (data missing from two patients)] and 28 days (IQR: 16–58 days; = 36), respectively. Disease progression was a commonly cited reason for hospitalization in patients receiving supportive care (48%).

Table 5. Resource utilization outside of clinical trials
OverallSystemic therapyaBSC
  1. BSC, best supportive care; NA, not available. aAs part of one line of treatment; btotal number of patients with data (i.e. missing are excluded). cPatient may be represented in more than one category; dMedical management agents used in ≥ 5% of patients; ePatients with supportive care only are considered as not having medical management due to treatment-related adverse events.

Patients, Nb15511568
Any hospitalization, n (%)43 (27·7)14 (12·2)31 (46)
Any hospice care, n (%)36 (23·2)13 (11·3)25 (37)
Any outpatient care, n (%)117 (75·5)80 (69·6)46 (68)
Patients requiring hospitalizationc, n441531
Reason for hospitalizationc, n (%)
Drug administration2 (4)1 (7)1 (3)
Disease progression16 (36)1 (7)15 (48)
Toxicity3 (7)3 (20)0
Other28 (64)10 (67)18 (58)
Medical management of treatment-related adverse eventsd, n2929NAe


These U.K.-specific data are from an observational and retrospective study performed across 31 sites in France, Italy and the U.K. that aimed to provide an overview of dominant treatment practices and clinical outcomes for patients with advanced melanoma attending a participating site between 1 July 2005 and 30 June 2006.

According to this U.K. analysis, approximately a quarter of patients initially presented with advanced melanoma during the inclusion period, possibly reflecting a bias towards higher-risk patients among the selected centres. Outside of clinical trials, most patients received dacarbazine as their first- or second-line treatment. Only three other agents (vindesine, temozolomide and carboplatin) were in widespread use, reflecting the limited treatment options available to patients at this time. Of note, approximately 10% of U.K. patients received immunotherapy. However, at the time of this study, use of interferon was largely driven by a single site in the U.K. and does not represent common practice across the U.K. Although this study did not report the types of immunotherapies used, examples of treatments investigated within clinical trials may have included interleukin-2 (given alone, after radiotherapy or in combination with other immunotherapies or chemotherapies) and antigen-based or dendritic cell-based cancer vaccines.

Response rates, disease control rates and OS for the U.K. population decreased with each successive line of treatment and responses were generally of short duration, although anecdotal evidence of patients with complete responses lasting more than 6 years following second-line systemic therapy was provided. However, it should be noted that data on response rates, PFS and durations of treatment benefit were not collected in a uniform way, are likely to reflect centre-specific confounding factors and may be biased by the frequency of local tumour assessments or response criteria used. For instance, reported response rates were higher than expected for the agents used but were derived from local reports and not subject to central review.[14]

The study indicates that patients were more likely to receive systemic therapy than supportive care alone and to be hospitalized if they were receiving supportive care. Inpatient hospitalization is generally the largest cost driver for any disease,[25] with the National Health Service currently spending four times as much money on the acute treatment of cancer than on community or palliative care.[26] This suggests that some form of active treatment may be the optimal choice for these patients in terms of both symptom control and resource utilization. Alternatively, U.K. oncologists may be correctly identifying patients in the terminal phase of melanoma and these patients have high supportive care requirements. Indeed, most patients receiving supportive care were hospitalized for reasons other than disease progression or treatment toxicity, suggesting high levels of symptom control. It should be noted that although case records were accumulated in accordance with a prespecified protocol, the quality and completeness of the study data were dependent on the level of detail and accuracy with which patient medical charts were maintained at participating sites. This may be a particularly relevant limitation with respect to collection of healthcare resource utilization data, as documentation of this information can vary from site to site. The economic impact of pan-European healthcare resource utilization from MELODY was reported by Johnston et al.[21]

MELODY has both strengths and limitations. It is the first multi-country observational study in advanced melanoma. Data were collected retrospectively in accordance with a protocol agreed by an independent ethics committee, rather than from the literature or anecdotal evidence. The minimum sample size was predetermined to facilitate accurate determination of treatment patterns. Sites were selected to be representative of melanoma management in France, Italy and the U.K., although many regions in the U.K. were not represented. Many sites are reluctant to take part in retrospective, observational studies such as MELODY, prioritizing participation in clinical trials or citing lack of resource or time constraints. In the U.K., for example, more than half of the initially identified sites declined further participation, mostly due to a lack of resources for performing the data collection in the expected time frame. An additional 25 sites were then selected to achieve the recruitment target of 10 sites. Overall, 20 sites declined or did not respond.

The data in this study are observational and retrospective and should be interpreted accordingly, particularly as nonstandardized reporting of data across the different sites is likely to have introduced information bias. For example, U.K.-specific confounding factors, such as centre-specific access to clinical trials, are not reported here. Additionally, data relating to hospice care are not available, making it difficult to conclude whether systemic therapy delays disease progression or alleviates symptoms more effectively than best supportive care. Furthermore, the lack of data on follow-up, including patients who stopped treatment prematurely, transferred out of the treatment facility or for whom death was not documented in the patient's medical chart (hospice data), is an inherent limitation of a retrospective study. However, loss to follow-up might be expected to be low, due to the short median duration of survival and frequency of contact between clinicians treating patients with advanced disease.

Survival figures in this analysis are higher than expected. Patients with metastatic melanoma have a median survival of 6–10 months,[27] whereas in this analysis, median OS among all patients was 15·6 months. The longer OS observed in U.K. patients and in the MELODY study population as a whole[22] may be a reflection of survival being defined from the date of diagnosis of advanced melanoma, rather than from the point of registration in a clinical trial.[28] Furthermore, the retrospective design of this study means that patients with a good prognosis who were still alive at the time of inclusion would contribute to the OS time, while those diagnosed at the same time who had died prior to the inclusion year would not be represented. To account for this potential selection bias, additional OS analyses were performed for patients first diagnosed with advanced disease during the study inclusion year alone, i.e. among incident rather than prevalent cases. In these 144 patients, median OS was reduced to 10·1 months, in line with previous observations (Table 4).[15, 29-31] Survival has also been shown to worsen as LDH increases.[32] A retrospective, single-institution analysis of 212 chemotherapy-treated patients with metastatic melanoma showed that patients with a normal LDH level survived twice as long as those with elevated LDH (12 vs. 6 months, respectively).[33] At the point of inclusion, twice as many MELODY U.K. patients had normal LDH than elevated LDH; therefore, it is possible that baseline characteristics contributed to the longer-than-expected-duration of survival, as discussed in the pan-European report of MELODY by Lebbe et al.[22] It should be noted that direct comparisons between observational studies and randomized controlled trials are not robust and should be interpreted with caution. Additionally, it is difficult to determine if metastatic staging was, in fact, a contributing factor to longer-than-expected survival, as M-stage classification data were missing from 117 of the 220 (53%) patients.

Two new treatments have been approved by the European Union for the treatment of advanced melanoma since this survey was conducted: vemurafenib and ipilimumab. Both agents have shown a survival benefit in pivotal trials.[5, 7] Accurate estimation of the cost-effectiveness of these treatments requires an understanding of treatment patterns and healthcare resource utilization in advanced melanoma. This is the first study to provide such specific U.K.-based data.

In conclusion, although the retrospective and observational nature of this study makes it difficult to draw robust conclusions, the high proportion of patients progressing to second-line therapy and/or entering clinical trials, and the ongoing reliance on dacarbazine, supports the clinical need for new treatment options in advanced melanoma. Additionally, systemic and palliative treatments used to manage advanced melanoma in the U.K. are associated with considerable healthcare resource utilization. These data highlight that healthcare resource utilization extends beyond primary treatments and that additional medical management for advanced melanoma is required.