You are here
Melanoma of unknown primary origin: A population-based study in the Netherlands
European Journal of Cancer, 3, 49, pages 676 - 683
Few population-based studies have been published on melanoma of unknown primary origin (MUP). This study’s aim is to describe characteristics and survival of MUP patients in the Netherlands, based on nationwide data from the Netherlands Cancer Registry (NCR).
Patient and tumour characteristics of MUP patients were retrieved from the NCR. Subgroups were made according to metastatic site: nodal or distant. Survival rates were calculated using the Kaplan–Meier method. To obtain a better insight in the composition and prognosis of the MUP group, the survival was compared to that of patients with melanoma of a known primary origin (MKP), tumour-node-metastasis (TNM) stage III and IV.
Of all 33,181 melanoma patients diagnosed between 2003 and 2009, 2.6% (n = 857) were diagnosed with MUP. MUP patients with nodal metastases had a similar survival as MKP stage III with macroscopic nodal involvement. After stratification according to the number of involved lymph nodes, the survival of patients with nodal metastases with one involved lymph node was not significantly different between MUP and MKP. The survival of MUP patients with two or more involved lymph nodes was slightly worse than that of MKP stage III patients with macroscopic nodal involvement with two or more involved lymph nodes. MUP patients with distant metastases had a similar survival as MKP stage IV. After stratification according to number of metastatic sites and metastatic site category, the survival in MKP stage IV patients with (sub)cutaneous metastases was slightly worse than MUP distant patients with (sub)cutaneous metastases.
The results of this study imply that MUP patients form a heterogeneous group, and that MUP patients with nodal metastases could be classified as stage III melanoma with macroscopic nodal involvement, and MUP patients with distant metastases as stage IV melanoma.
Keywords: Melanoma, Neoplasms, Unknown primary, Survival analysis, Epidemiology.
Melanoma is usually discovered as a skin lesion. However, about 2–3% of all melanoma patients present themselves with a melanoma metastasis without a detectable primary tumour.1, 2, and 3 This is referred to as melanoma of unknown primary origin (MUP). Several theories have been proposed for the aetiology of MUP, including the following: a MUP may arise from (i) a melanoma that was excised and misdiagnosed or not further investigated histopathologically; (ii) a regressed melanoma; or (iii) malignant transformation of a naevus cell in a lymph node or other non-skin tissue. In support of the last two theories, partial spontaneous regression of melanomas occurs frequently and may be due to immunological mechanisms.4 and 5 Benign naevus cells are commonly found in lymph nodes and other tissues, and melanomas arising from naevus cells in lymph nodes have been described.6 and 7
MUP patients comprise a heterogeneous group. Like patients with metastatic melanoma of a known primary (MKP), MUP patients can present with (sub)cutaneous, nodal metastasis and/or metastasis to the viscera, bones or brain. In patients with only (sub)cutaneous or nodal metastases it is impossible to differentiate between regional versus distant metastases, as the primary site is unknown. This distinction would be important for the assessment of prognosis.
Most of what is known about the prognosis of MUP patients is derived from hospital-based studies. Based on these studies, MUP patients with nodal metastases seem to have a similar8, 9, and 10 to better11, 12, and 13 survival than MKP patients with nodal metastases. MUP patients with distant metastases seem to have a survival similar or even better compared with MKP patients with distant metastases.1, 14, and 15
Population-based studies on this subject are scarce. In 1998, Chang et al. reported on 84,836 melanoma cases diagnosed between 1985 and 1994 from the American National Cancer Data Base, of which 2.2% had an unknown primary. 1 To our knowledge, no recent population-based study on MUP has been published since. This type of study can provide interesting, unselected information on this patient group. Therefore, the aim of this study is to describe characteristics and survival of the different subgroups of MUP patients, based on unselected, population-based data retrieved from the Netherlands Cancer Registry (NCR) database. To give insight in the composition and prognosis of this heterogeneous patient group, survival analyses were performed and stratified comparisons with stage III and IV MKP patients were made.
2. Patients and methods
This population-based study was based on all melanoma patients diagnosed in the Netherlands between 1st January 2003 and 31th December 2009. In the NCR, held by the Comprehensive Cancer Centre the Netherlands, all newly diagnosed malignancies are recorded. It has nationwide coverage since 1989. Patient and tumour characteristics, primary treatment and vital status are recorded in the NCR. Vital status of all patients registered in the NCR is updated annually by record linkage to the Dutch Municipal Personal Records Database, which keeps information about vital status of all inhabitants in the Netherlands. For patients diagnosed with metastatic cancer, the NCR records the location of metastases using the International Classification of Diseases-Oncology (ICD-O) classification.
Patients diagnosed with melanoma were identified by the morphology codes 8720-8780 (ICD-O-3). Based on the topography code, these patients were classified as MUP patients (topography C80) or MKP patients (all topography codes except C80). Patients with a C80 topography but with a prior history of primary melanoma were excluded from the analyses. Because cutaneous and non-cutaneous melanomas comprise separate entities with different tumour-node-metastasis (TNM) staging classifications, only MKP patients with a cutaneous melanoma (topography C44) were included in this study. All non-cutaneous primary melanoma localisations were excluded. Since the Union for International Cancer Control (UICC) TNM staging protocol for cutaneous melanoma changed markedly in the transition from the fifth to the sixth edition, only cases diagnosed after the implementation of the sixth edition were included (1st January 2003).16, 17, and 18 For each patient, age at diagnosis, gender, clinical and post-surgical TNM stage, number of involved lymph nodes, location of metastases present at diagnosis, primary treatment and vital status were retrieved from the NCR.
Two MUP subgroups were made based on the localisation of their metastases. Subgroups were made based on the most stage-defining metastasis: (i) only nodal metastases and (ii) any (sub)cutaneous, visceral, bone or brain metastasis, hereafter called MUP nodal and MUP distant respectively. Patients with cutaneous or subcutaneous metastases were categorised as ‘distant’, because most patients with subcutaneous metastases likely have distant metastases. Patients with cutaneous metastases could in fact have in transit or satellite metastases and therefore could have a better prognosis. Unfortunately a distinction between locoregional and distant skin metastases cannot be made, since the primary tumour is unknown. We therefore chose this classification, and analysed patients with (sub)cutaneous separately in the stratified analysis. In 142 MUP patients (17%) information on the metastatic site was missing. These patients were excluded from the subgroup analyses. MKP patients were divided into TNM stage III and IV. MUP nodal patients most likely have macroscopic clinically palpable metastases, and MKP stage III also includes patients with microscopic nodal metastases and in transit or satellite metastases. Therefore, MKP stage III was further subdivided into patients with microscopic nodal involvement, macroscopic nodal involvement, and in transit or satellites metastases. For further comparisons with the MUP nodal group, only MKP stage III patients with macroscopic metastases were taken into account, hereafter called MKP stage III macroscopic. For the MUP nodal and the MKP stage III macroscopic group, the number of involved lymph nodes was recorded as zero, one or two or more. This classification was based on available information on the number of lymph nodes from the NCR. The number of distant metastatic sites (one, or two or more) and three major metastatic site categories ((sub)cutaneous or lymph nodes beyond regional lymph nodes, lung and other sites) were recorded for the MUP distant and MKP stage IV subgroups.
2.2. Statistical methods
The percentage of MUP relative to all melanoma cases diagnosed between 2003 and 2009 was determined. Crude (i.e. all cause) survival was calculated using the Kaplan–Meier method. Follow-up time was calculated and defined as the time between date of diagnosis and date of death, date of emigration or end of follow-up (1st January 2010), whichever came first. Two- and five-year survival estimates with 95% confidence intervals (95% CI), and median survival time (months) were estimated for the two MUP subgroups and MKP stage III and IV. In MUP nodal and MKP stage III macroscopic patients, the survival estimates were calculated for one versus two or more involved lymph nodes, and in the MUP distant and MKP stage IV patients, for one versus two or more metastatic sites, and for the metastatic site categories separately. Differences in the short term (5-year) survival were tested with the Breslow (Generalised Wilcoxon) test. A p-value of 0.05 was considered statistically significant.
In order to evaluate the comparability of the prognosis of the different MUP subgroups with the prognosis of the MKP stage groups, the survival of MUP patients with nodal metastases was compared with the survival of MKP stage III macroscopic patients, and the survival of MUP patients with distant metastasis was compared to that of MKP stage IV patients. In addition, for the MUP nodal and MKP stage III macroscopic groups, stratified comparisons were made according to the number of involved lymph nodes. This was also done for the MUP distant and MKP stage IV groups, according to the number of metastatic sites, and according to metastatic site category. Gender specific survival was calculated and compared for all MUP and MKP subgroups.
The statistical analyses were performed with SPSS 18.0 software (SPSS Inc., Chicago, IL). The data retrieval was approved by the Institutional Review Board (IRB) of the Netherlands Cancer Registry.
Overall, 33,181 patients were diagnosed with melanoma between 2003 and 2009. Of all melanoma patients, 2.6% (n = 857) were diagnosed with a melanoma of unknown primary origin.
In Table 1 patient and tumour characteristics of MUP patients by subgroup and cutaneous MKP stage III and IV patients are presented. Of the MUP patients, 28% (n = 237) had only nodal metastases and 56% (n = 478) had distant metastasis. For 17% (n = 142) the metastatic site was unknown. The male:female ratio was not significantly different between the subgroups of MUP and MKP patients, as was median age. The proportion of patients that was treated with a lymph node dissection was not significantly different between MUP nodal and MKP stage III macroscopic patients (60% versus 61%, p = 0.935), nor was the proportion of patients that received radiotherapy (20% versus 16%, p = 0.169) or systemic therapy (29% versus 27%, p = 0.635) between MUP distant and MKP stage IV.
|Subcategory||MUP (n = 857)||MKP (n = 1975)|
|Nodal||Distant a||Stage III||Stage IV|
|n (%)||n (%)||n (%)||n (%)|
|Total no. b||237 (27.7)||478 (55.8)||1689 (85.5)||286 (14.5)|
|Male||138 (58.2)||270 (56.5)||931 (55.1)||166 (58.0)|
|Female||99 (41.8)||208 (43.5)||758 (44.9)||120 (42.0)|
|Age at diagnosis, median (range)||61 (15–92)||61 (7–97)||57 (8–99)||61 (18–97)|
|Stage III substage|
|Microscopic nodal||NA||NA||783 (46.4)||NA|
|Macroscopic nodal||625 (37.0)|
|In transit or satellite||281 (16.6)|
|Lymph node dissection c||143 (60.3)||379 (60.6)|
|Systemic therapy||138 (28.9)||78 (27.3)|
|Radiotherapy||96 (20.1)||46 (16.1)|
|Number of nodal metastases|
|1||214 (90.3)||806 (47.7)|
|⩾2||23 (9.7)||540 (32.0)|
|Number of distant metastatic sites|
|1||NA||244 (51.0)||NA||107 (37.4)|
|⩾2||234 (49.0)||127 (44.4)|
|Site of distant metastases|
|(Sub)cutaneous or LN beyond regional LN d||NA||106 (12.4)||NA||47 (16.4)|
|Lung||41 (4.7)||43 (15)|
|Other sites||331 (38.7)||144 (81.8)|
a MUP distant includes patients with any (sub)cutaneous, visceral, bone or brain metastasis.
b For 142 MUP patients (16.6%), the metastatic site was unknown or missing.
c In MKP patients, the number of lymph node dissections is reported for MKP stage III patients with macroscopic nodal involvement.
d In MUP patients, this includes only patients with (sub)cutaneous metastases, since patients with nodal metastases were categorised into a separate group.
Abbreviations: MUP, melanoma of unknown primary origin; MKP, melanoma of known primary origin; NA, not available; LN, lymph node.
Survival curves of the MUP nodal and distant subgroups are presented in Fig. 1 . The survival of the nodal subgroup was significantly different from that of the distant subgroup (p < 0.001).
Fig. 2 A shows a significantly worse survival of the MUP nodal subgroup compared with MKP stage III (p = 0.001), though a similar survival of the MUP nodal subgroup compared to MKP stage III macroscopic (p = 0.374). Furthermore, Fig. 2 B shows a similar survival of the MUP distant subgroup compared with MKP stage IV (p = 0.486). In Fig. 3 A and B, stratified comparisons are shown for the survival of the MUP nodal and MKP stage III macroscopic subgroups, according to the number of involved lymph nodes. MUP nodal patients with one involved lymph node had a similar survival as MKP stage III macroscopic patients with one involved lymph node (p = 0.632). However, MUP nodal patients with two or more lymph nodes involved had a worse survival than MKP stage III macroscopic patients with two or more involved lymph nodes (p < 0.001). Fig. 4 A and B show stratified comparisons for the survival of the MUP distant and MKP stage IV subgroups, according to the number of metastatic sites. MUP distant patients and MKP stage IV patients with one metastatic site had a similar survival (p = 0.211). MUP distant patients with two or more metastatic sites had a slightly worse survival than MKP stage IV patients with two or more metastatic sites (p = 0.008). Fig. 5 A–C show stratified comparisons for the survival of the MUP distant and MKP stage IV groups, according to metastatic site category. MUP distant patients with (sub)cutaneous metastases had a better survival than MKP stage IV patients with (sub)cutaneous or distant lymph nodes (p = 0.036). MUP distant and MKP stage IV patients with lung metastases had a similar survival (p = 0.800), as well as patients with other distant metastases (p = 0.233).
In Table 2 , 2- and 5-year survival estimates with 95% CI and median survival time (months) are presented for all the subgroups previously mentioned.
|N||Two-year survival (% [95% CI])||Five-year survival (% [95% CI])||Median survival time (months)|
|Nodal, all||327||64.6 (57.9–71.4)||52.0 (43.9–60.0)||NA|
|Nodal, 1 LN involved||214||68.8 (61.9–75.8)||56.0 (47.6–64.4)||NA|
|Nodal, ⩾2 LN involved||23||22.7 (2.3–43.1)||15.1 (0–33.3)||8.9|
|Distant, all||478||23.9 (19.7–28.2)||14.0 (8.7–19.3)||6.7|
|Distant, 1 metastatic site||244||42.6 (35.7–49.6)||25.5 (16.1–34.8)||16.8|
|Distant, ⩾2 metastatic sites||234||4.2 (1.2–7.2)||0||3.2|
|Distant, (sub)cutaneous||106||62.2 (51.6–72.9)||35.3 (18.5–52.0)||54.7|
|Distant, lung||41||36.6 (20.4–52.1)||14.7 (0–30.7)||12.5|
|Distant, other sites||331||10.0 (6.2–13.7)||7.7 (3.7–11.7)||3.2|
|Stage III, all||1689||75.8 (73.5–78.1)||54.6 (51.3–57.9)||NA|
|Stage III, macro, all||625||52.0 (48.2–55.8)||35.0 (30.7–39.3)||25.6|
|Stage III, macro, 1 LN involved||271||67.9 (62.1–73.6)||50.1 (43.0–57.3)||NA|
|Stage III, macro, ⩾2 LN involved||306||46.7 (41.0–52.4)||27.1 (20.7–33.5)||21.4|
|Stage IV, all||286||20.1 (15.0–25.1)||8.9 (4.5–13.3)||6.3|
|Stage IV, 1 metastatic site||107||33.6 (23.7–43.4)||14.5 (4.8–24.3)||11.9|
|Stage IV, ⩾2 metastatic sites||127||6.8 (1.7–11.8)||4.1 (0–8.3)||4.4|
|Stage IV, subcutaneous or LN beyond regional LN||47||44.9 (28.5–61.3)||35.3 (17.5–53.0)||17.3|
|Stage IV, lung||43||34.7 (19.6–49.7)||5.4 (0–15.3)||12.4|
|Stage IV, other sites||144||5.5 (0–9.9)||1.5 (1.2–4.1)||4.4|
Abbreviations: MUP, melanoma of unknown primary origin; MKP, melanoma of known primary origin; NA, not available; LN, lymph node; macro, macroscopic nodal involvement.
Gender specific survival analyses showed no significant differences between males and females in the MUP subgroups or the MKP stage III and IV groups.
When comparing the survival of MUP patients with nodal metastases to that of MKP stage III, we found that the latter fared slightly though significantly better. However, when we only took MKP stage III patients with macroscopic nodal involvement into account, the survival was not significantly different. This can be explained by the fact that most MUP nodal patients probably present with macroscopic disease. After stratification according to the number of involved lymph nodes, the survival was worse in the MUP nodal subgroup with two or more involved lymph nodes. This may be explained by the fact that in the MUP nodal subgroup, patients with regional metastases are probably intermixed with patients with metastases to distant nodal basins. Most likely, patients with two or more lymph nodes involved belong in this latter category.
It can be assumed that MUP nodal patients have a similar survival as stage III patients with macroscopic nodal involvement, and therefore could be regarded as TNM stage III patients with macroscopic involvement. However, when MUP patients have two or more lymph nodes involved, their prognosis may be worse. Since this stratified analysis was based on only 25 patients in the MUP nodal subgroup, conclusions should be drawn carefully.
Other studies reported a similar to better survival for MUP nodal patients,8, 9, 10, 11, 12, 15, 19, and 20 and Pfeil et al. recently posed that MUP patients with lymph node metastases can be staged as AJCC stage III. 3 However, based on our results, we conclude that MUP nodal patients can only be compared with MKP stage III patients with macroscopic nodal involvement. Furthermore, the MUP nodal group theoretically also includes patients with distant nodal metastases.
In this study MUP distant patients appear to have a similar survival as MKP stage IV patients. When making stratified comparisons the survival was not significantly different in patients with one metastatic site and patients with lung or other metastatic sites. Only in the patients with (sub)cutaneous metastases the survival was significantly better for MUP patients. An explanation might be that MUP patients with (sub)cutaneous metastases could also have in transit or satellite metastases, with a better prognosis. In patients with two or more metastatic sites, the survival was significantly worse for MUP patients. This survival difference, however, was small. We conclude therefore that MUP distant patients could be staged as TNM stage IV. Only a few other recent studies reported on MUP patients with distant metastases, and some showed a similar survival to MKP stage IV patients,3 and 10 including the population-based report by Chang et al. which reported a 5-year survival of 16% compared with 18% in the stage IV MKP group. 1
Only one study by Lee et al. made detailed comparisons between the survival of stage IV categories M1a, M1b and M1c of MUP and MKP patients, and they found a better survival for the MUP stage IV M1b and M1c group (i.e. metastases to the lungs or other viscera) compared with MKP. 14 We could not replicate this survival benefit for the MUP distant group over the MKP stage IV group with lung or other metastases.
It is argued by several researchers that MUP patients have a better prognosis than MKP patients due to the same immunological mechanisms responsible for the regression of the primary tumour.12, 13, and 15 The literature on the survival of MUP patients versus MKP patients however, is not consistent. Moreover, Guiliano et al. were unable to identify differences in cellular immune responses between patients with a known or unknown primary melanoma. 4
In this study, females had a better survival than males in both the MUP and MKP subgroups (data not shown), which is in concordance with findings from previous studies.13 and 21 It is suggested that melanomas in women have a lower propensity to metastasize, which might explain this observed survival difference. 22 This, in addition to the similar male:female ratio and median age in both groups, gives another indication that the MUP patient group has a similar composition as the MKP group.
When comparing the absolute survival estimates of the different MUP subgroups (i.e. nodal and distant) to those found by other studies, the absolute 5-year survival for MUP patients with nodal metastases (52%) was similar to that found in other study populations (46–55%).10, 11, and 12 In some studies it was worse than in the present study (39–41%)2, 9, and 13 and in some studies a better survival was reported (57–60%).3, 19, and 20
For MUP patients with distant metastases, most studies found a better survival than in the present study.3, 14, and 20 These studies were hospital-based, thereby possibly selecting patients from a certain region or income group or even prognosis. Some only included melanoma patients with resected nodal metastases, possibly selecting a healthier group with a survival benefit over other, unselected patient groups. In the population-based study from the United States by Chang and colleagues, the 5-year survival of MUP patients with nodal metastasis was slightly lower (46% versus 52% in our cohort), though that of MUP patients with distant metastases was better (16% versus 14% in our cohort). 1 This study was performed in the United States, so the population under study may have had a different ethnic and socio-economical composition than the Dutch population.
In this study we were able to use data of all Dutch melanoma patients diagnosed from 2003 to 2009. This gave us the opportunity to identify the actual proportion of MUP relative to all melanoma diagnoses and review a large, unselected group of MUP patients. Furthermore, we had information on the localisation of metastases in most MUP patients, making stratified analyses possible. To the best of our knowledge, this is only the second population-based study that describes the characteristics and survival of MUP.
As the data was derived from the NCR, only information recorded in the NCR could be included in this study. Prognostic factors such as tumour ulceration, micro- or macrometastasis, or elevated LDH were not available, or only available for MKP and not for MUP patients. Information on the primary treatment was available, and we found no significant differences between the MUP and MKP groups. We decided to perform stratified comparisons for the number of nodes and metastatic sites, and metastatic site categories, to give a descriptive analysis of the composition and survival of the MUP group. A Cox regression model including more prognostic factors would make it possible to explain survival differences better, but unfortunately the data were limited.
In conclusion, the results of this study imply that MUP patients represent a heterogeneous group, that MUP patients with nodal metastases have a similar survival compared with MKP stage III patients with macroscopic involvement and that MUP patients with distant metastases have a similar survival as MKP stage IV patients. The problem remains that physicians cannot distinguish which of their MUP patients have a regional or a distant (sub)cutaneous or nodal metastasis. They should mention this fact when informing these patients about their prognosis and should take this into account when classifying them into the existing TNM staging categories.
Conflict of interest statement
Role of the funding sources
Dutch Cancer Society, grant number KUN2010-4616: no involvement in the study design, collection, analysis and interpretation of the data, in the writing of the manuscript or in the decision to submit the manuscript for publication.
Radboud University Medical Centre, Junior Researcher grant: no involvement in the study design, collection, analysis and interpretation of the data, in the writing of the manuscript or in the decision to submit the manuscript for publication.
- 1 A.E. Chang, L.H. Karnell, H.R. Menck. The National Cancer Data Base report on cutaneous and noncutaneous melanoma: a summary of 84,836 cases from the past decade. The American College of Surgeons Commission on Cancer and the American Cancer Society. Cancer. 1998;83:1664-1678
- 2 K.A. Katz, E. Jonasch, F.S. Hodi, et al. Melanoma of unknown primary: experience at Massachusetts General Hospital and Dana-Farber Cancer Institute. Melanoma Res. 2005;15:77-82
- 3 A.F. Pfeil, U. Leiter, P.G. Buettner, et al. Melanoma of unknown primary is correctly classified by the AJCC melanoma classification from 2009. Melanoma Res. 2011;21:228-234
- 4 A.E. Giuliano, H.S. Moseley, R.F. Irie, S.H. Golub, D.L. Morton. Immunologic aspects of unknown primary melanoma. Surgery. 1980;87:101-105
- 5 V.J. McGovern. Spontaneous regression of melanoma. Pathology. 1975;7:91-99
- 6 T. Dasgupta, L. Bowden, J.W. Berg. Malignant melanoma of unknown primary origin. Surg Gynecol Obstet. 1963;117:341-345
- 7 S.W. McCarthy, A.A. Palmer, P.M. Bale, E. Hirst. Naevus cells in lymph nodes. Pathology. 1974;6:351-358
- 8 F. Laveau, M.C. Picot, O. Dereure, J.J. Guilhou, B. Guillot. Metastatic melanoma of unknown primary site. Ann Dermatol Venereol. 2001;128:893-898
- 9 P. Rutkowski, Z.I. Nowecki, W. Dziewirski, et al. Melanoma without a detectable primary site with metastases to lymph nodes. Dermatol Surg. 2010;36:868-876
- 10 B. Schlagenhauff, W. Stroebel, U. Ellwanger, et al. Metastatic melanoma of unknown primary origin shows prognostic similarities to regional metastatic melanoma: recommendations for initial staging examinations. Cancer. 1997;80:60-65
- 11 J.N. Cormier, Y. Xing, L. Feng, et al. Metastatic melanoma to lymph nodes in patients with unknown primary sites. Cancer. 2006;106:2012-2020
- 12 C.C. Lee, M.B. Faries, L.A. Wanek, D.L. Morton. Improved survival after lymphadenectomy for nodal metastasis from an unknown primary melanoma. J Clin Oncol. 2008;26:535-541
- 13 S.P. Prens, A.P. van der Ploeg, A.C. van Akkooi, et al. Outcome after therapeutic lymph node dissection in patients with unknown primary melanoma site. Ann Surg Oncol. 2011;18:3586-3592
- 14 C.C. Lee, M.B. Faries, L.A. Wanek, D.L. Morton. Improved survival for stage IV melanoma from an unknown primary site. J Clin Oncol. 2009;27:3489-3495
- 15 G. Vijuk, A.S. Coates. Survival of patients with visceral metastatic melanoma from an occult primary lesion: a retrospective matched cohort study. Ann Oncol. 1998;9:419-422
- 16 L.H. Sobin. TNM, sixth edition: new developments in general concepts and rules. Semin Surg Oncol. 2003;21:19-22
- 17 L.H. Sobin, C. Wittekind, International Union Against Cancer. TNM classification of malignant tumours. 5th ed. (John Wiley, New York, 1997) xxiii, 227 p.
- 18 L.H. Sobin, C. Wittekind, International Union Against Cancer. TNM: classification of malignant tumours. 6th ed. (Wiley–Liss, New York, 2002) xxiii, 239 p.
- 19 K.K. Anbari, L.M. Schuchter, L.P. Bucky, et al. Melanoma of unknown primary site: presentation, treatment, and prognosis – a single institution study. University of Pennsylvania Pigmented Lesion Study Group. Cancer. 1997;79:1816-1821
- 20 P. Savoia, P. Fava, S. Osella-Abate, et al. Melanoma of unknown primary site: a 33-year experience at the Turin Melanoma Centre. Melanoma Res. 2010;20:227-232
- 21 E. de Vries, T.E. Nijsten, O. Visser, et al. Superior survival of females among 10,538 Dutch melanoma patients is independent of Breslow thickness, histologic type and tumor site. Ann Oncol. 2008;19:583-589
- 22 A. Joosse, E. de Vries, R. Eckel, et al. Gender differences in melanoma survival: female patients have a decreased risk of metastasis. J Invest Dermatol. 2011;131:719-726
a Department of Dermatology, Radboud University Medical Centre, 370, P.O. Box 9101, NL-6500 HB Nijmegen, The Netherlands
b Department of Epidemiology, Biostatistics and HTA, Radboud University Medical Centre, 133, P.O. Box 9101, NL-6500 HB Nijmegen, The Netherlands
c Comprehensive Cancer Centre the Netherlands, P.O. Box 19079, NL-3501 DB Utrecht, The Netherlands
d Department of Urology, Radboud University Medical Centre, 659, P.O. Box 9101, NL-6500 HB Nijmegen, The Netherlands
© 2012 Elsevier Ltd, All rights reserved.