Factors related to the presentation of thin and thick nodular melanoma from a population-based cancer registry in Queensland Australia


  • Alan C. Geller MPH, RN,

    Corresponding author
    1. Division of Public Health Practice, Harvard School of Public Health, Boston, Massachusetts
    • Harvard School of Public Health, Division of Public Health Practice, Landmark Center, 401 Park Drive, 3 East, Boston, MA 02215
    Search for more papers by this author
    • Fax: (617) 495-8543

  • Mark Elwood MD,

    1. British Columbia Cancer Agency, Vancouver, Canada
    Search for more papers by this author
  • Susan M. Swetter MD,

    1. Department of Dermatology, Pigmented Lesion and Melanoma Program, Veterans Affairs Palo Alto Health Care System, Palo Alto, California
    2. Stanford University Medical Center, Stanford, California
    Search for more papers by this author
  • Daniel R. Brooks ScD, MPH,

    1. Department of Epidemiology, Boston University School of Public Health, Boston, Massachusetts
    Search for more papers by this author
  • Joanne Aitken PhD,

    1. Viertel Centre for Research in Cancer Control, Cancer Council Queensland, Brisbane, Australia
    2. School of Population Health, University of Queensland, Brisbane, Australia
    Search for more papers by this author
  • Philippa H. Youl MPH,

    1. Viertel Centre for Research in Cancer Control, Cancer Council Queensland, Brisbane, Australia
    Search for more papers by this author
  • Marie-France Demierre MD,

    1. Department of Dermatology, Boston University School of Medicine, Boston, Massachusetts
    Search for more papers by this author
  • Peter D. Baade PhD

    1. Viertel Centre for Research in Cancer Control, Cancer Council Queensland, Brisbane, Australia
    2. School of Public Health, Queensland University of Technology, Brisbane, Australia
    Search for more papers by this author



Worldwide, the incidence of thick melanoma has not declined, and the nodular melanoma (NM) subtype accounts for nearly 40% of newly diagnosed thick melanoma. To assess differences between patients with thin (≤2.00 mm) and thick (≥2.01 mm) nodular melanoma, the authors evaluated factors such as demographics, melanoma detection patterns, tumor visibility, and physician screening for NM alone and compared clinical presentation and anatomic location of NM with superficial spreading melanoma (SSM).


The authors used data from a large population-based study of Queensland (Australia) residents diagnosed with melanoma. Queensland residents aged 20 to 75 years with histologically confirmed first primary invasive cutaneous melanoma were eligible for the study, and all questionnaires were conducted by telephone (response rate, 77.9%).


During this 4-year period, 369 patients with nodular melanoma were interviewed, of whom 56.7% were diagnosed with tumors ≤2.00 mm. Men, older individuals, and those who had not been screened by a physician in the past 3 years were more likely to have nodular tumors of greater thickness. Thickest nodular melanoma (4 mm+) was also most common in persons who had not been screened by a physician within the past 3 years (odds ratio, 3.75; 95% confidence interval, 1.47-9.59). Forty-six percent of patients with thin nodular melanoma (≤2.00 mm) reported a change in color, compared with 64% of patients with thin SSM and 26% of patients with thick nodular melanoma (>2.00 mm).


Awareness of factors related to earlier detection of potentially fatal nodular melanomas, including the benefits of a physician examination, should be useful in enhancing public and professional education strategies. Particular awareness of clinical warning signs associated with thin nodular melanoma should allow for more prompt diagnosis and treatment of this subtype. Cancer 2009. © 2009 American Cancer Society.

Melanoma survival is strongly associated with thickness of the primary lesion.1 The median tumor thickness of cutaneous melanoma has decreased significantly over the past 3 decades as a result of earlier detection that has coincided with sharply rising incidence, particularly of thin melanoma, in Europe, Australia, and the United States.2-4 Almost 70% of these melanomas are of the superficial spreading histological subtype, and early detection is likely related to public and professional awareness of the ABCDE early melanoma warning signs.5

Despite the decrease in median tumor thickness, the incidence of thick melanoma has not declined.6-8 The nodular melanoma (NM) subtype accounts for a substantial fraction of newly diagnosed thick primary tumors with known histology in both Australia and the United States (Queensland Cancer Registry, personal communication).9-11 For example, in the US Surveillance, Epidemiology, and End Results (SEER) registry, NM comprises only 10% of all invasive melanoma, but nearly 50% of all melanoma >2.0 mm, a thickness cutoff point frequently used to distinguish thin from thick melanoma.10 In Queensland, NM comprises 8% of all invasive melanoma but 38% of melanoma thicker than 2.0 mm (Queensland Cancer Registry, personal communication). Similarly, a Swedish report found nodular melanoma to comprise 9% of all melanoma and 40% of all lesions >2.0 mm,12 and in Italy, NM makes up 65% of >2 mm melanomas.13 In Italy, from 1985-1987 to 1995-1997, there was a significant shift toward thinner melanoma, but most recent data showed that the nodular subtype and advanced age were the only variables significantly associated with thick melanoma.2

Nodular melanoma is commonly believed to be a rapidly growing subtype that eludes early detection because of its greater biological aggressiveness.11 New approaches to screening and education are required to reduce the incidence of thick melanoma. However, there are few reports that have characterized factors associated with differences between thin and thick nodular melanomas, and the diagnostic features of early NM have yet to be clearly defined.11 Therefore, we aimed to: 1) evaluate social demographics, discovery patterns, and phenotypic factors that differ between thin and thick NM; and 2) determine whether there were differences in clinical presentation between patients with thin NM, thick NM, thin SSM, and thick SSM. Awareness of factors related to earlier detection of the most commonly fatal melanoma subtype should be useful in enhancing new public and professional education strategies and could help to elucidate the interactions of tumor biology and clinical presentation of this subtype.


We used data from a large population-based study of Queensland (Australia) residents diagnosed with melanoma, for which the fieldwork methodology has been previously described in detail.14 Briefly, Queensland residents aged 20 to 75 years with histologically confirmed first primary invasive cutaneous melanoma between January 1, 2000 and December 31, 2003 were eligible for the study. For sampling and cost efficiency, a random 60% sample of participants with melanoma <0.75 mm were selected, and a census of those with greater tumor thickness were included. Patients with metastatic disease, a previous melanoma, or acral lentiginous or mucosal melanoma were excluded. After permission was obtained from the treating physician, patients were invited by letter to participate in a telephone survey. Ethical clearance for the original fieldwork was obtained from the University of Queensland Ethical Review Committee, while additional approval to conduct this subsequent analysis was obtained from Boston University.

Of 4839 eligible patients, physician consent was obtained for 4510 (93.2%); 3887 patients (80.3% of the total eligible sample) agreed to participate, and 3772 (77.9%) completed an interview. The median time between diagnosis and interview was 5 months (range, 1-26 months). Five people died before being interviewed. Of 542 eligible patients diagnosed with NM, physician consent was obtained for 512 (94.5%), 421 patients (77.7% of the total eligible sample) agreed to participate, and 369 (68.1%) completed an interview.

Data were collected from respondents via telephone by trained interviewers using a computer-assisted telephone interview system. Interviewers were blinded to specific tumor thickness and to histologic subtypes. Among other items, the interview collected information on sociodemographics (age in 10-year age groups, sex, and highest educational achievement categorized into primary, grades 7-10, or grade >11), anatomic site of melanoma, whether the melanoma was first noticed through a deliberate skin check, or noticed accidentally by the patient, partner, or physician, and performance of self-screening in the previous 3 years. Respondents were surveyed regarding physician screening (defined as whether a physician deliberately checked all or nearly all of the patient's body for the early signs of skin cancer during the last 3 years), self-reported “moliness” (according to diagrams illustrating many/some vs few/none), visibility of lesion (nonvisible vs visible), and skin reaction to summer sun (always/usually burns vs sometimes/never burns). Lesions were determined to be nonvisible to the patient if they were on the scalp, back, buttocks, back of the neck, ears, shoulders, and legs, and soles of the feet. All other melanomas were considered to be on visible sites of the body.

Respondents were asked “on the date you first believed something was wrong, what was it about the spot that made you or someone else believe something might be wrong?” Choices included: change in color, change in size, irregular shape, irritation/itch, lumpy/raised appearance, different from other spots, bleeding/weeping, dry/scaly, blister-like, pain, and pigmentation (brown, black, red/pink, pale/no color, gray/blue). Respondents were also asked if the lesion “just appeared.” Patients whose melanoma was detected by a physician were not asked this series of questions.

Further detail on the survey questions have been provided earlier.14, 15 The test-retest reliability of the survey was examined by reinterviewing 186 patients within 1 to 3 months of the first interview. Agreement ranged from 76.6% for 1 of the symptoms to 93% for skin self-examination.16

Interview data were combined with pathology data in the Queensland Cancer Registry, including melanoma thickness, histologic type, anatomic site of lesion (categorized as scalp and neck, upper extremities, lower extremities, back of trunk, and front on trunk), and level of invasion.

Statistical Analysis

We conducted bivariate and multivariate analyses of the associations between melanoma thickness and selected variables for the 369 respondents with histologically confirmed nodular melanomas. In all analyses, the proportion of thin melanoma was weighted to account for the undersampling (60%) of melanomas <0.75 mm in this study, in contrast to the inclusion of all thicker (≥0.75 mm) tumors.

Melanoma thickness was first treated as a continuous variable and modeled on a log scale. Associations are reported as ratios of the geometric means compared with the reference category with 95% confidence intervals. We report unadjusted bivariate estimates and multivariate associations with adjusted estimates for all of the explanatory variables shown in Table 1. Likelihood ratio (chi-square) statistics were used to assess the significance of variables from the multivariate models. Results are expressed as the exponential of the relevant parameter estimate from the model, and reflect the ratio of the geometric mean relative to the baseline category.

Table 1. Ratios of Geometric Means for Melanoma Thickness (Continuous Variable*) and Selected Variables (Nodular Melanomas Only, n = 369)
Estimate95% Confidence IntervalEstimate§95% Confidence Interval
  • *

    Melanoma thickness was modeled on a log scale, and treated as a continuous variable.

  • Estimates are the exponential of the relevant parameter estimate, and reflect the ratio of the geometric mean relative to the baseline.

  • Unadjusted estimates.

  • §

    Adjusted estimates are adjusted for all the other variables in the table.

  • Likelihood ratio statistics (type III) for significance of variable.

SexChi-square = 2.58, df = 1, P = .108Chi-square = 4.42, df = 1, P = .036
 Men (n = 222)1.130.97-1.311.171.01-1.35
 Women (n = 147)1.00 1.00 
Age groupChi-square = 34.82, df = 5, P < .001Chi-square = 27.27, df = 5, P < .001
 20-29 y (n = 13)0.980.66-1.451.030.70-1.52
 30-39 y (n = 39)1.030.79-1.331.140.89-1.47
 40-49 y (n = 74)1.00 1.00 
 50-59 y (n = 90)1.371.12-1.681.431.17-1.75
 60-69 y (n = 88)1.371.12-1.681.381.12-1.70
 70-75 y (n = 65)1.831.47-2.301.791.42-2.26
Highest educationChi-square = 14.34, df = 2, P < .001Chi-square = 4.20, df = 2, P = .123
 Primary (n = 35)1.631.26-2.101.311.01-1.69
 Years 7-10 (n = 128)1.130.97-1.321.060.91-1.23
 Years 11-12 and upward (n = 205)1.00 1.00 
Physician screen in previous 3 yearsChi-square = 7.43, df = 1, P = .006Chi-square = 5.10, df = 1, P = .024
 Screened (n = 103)1.00 1.00 
 Not screened (n = 266)1.251.06-1.461.201.02-1.41
Visibility of lesionChi-square = 4.57, df = 1, P = .033Chi-square = 6.68, df = 1, P = .010
 Not visible (n = 173)1.00 1.00 
 Visible (n = 196)1.171.01-1.351.201.05-1.38
Who first noticed the melanoma?Chi-square = 5.88, df = 2, P = .053Chi-square = 3.79, df = 2, P = .150
 Your partner (n = 99)1.120.87-1.441.120.85-1.48
 Yourself (n = 228)1.281.02-1.611.250.96-1.63
 Physician (n = 42)1.00 1.00 
Method of detectionChi-square = 3.39, df = 1, P = .066Chi-square = 0.28, df = 1, P = .597
 By accident (n = 325)1.230.99-1.541.070.83-1.39
 Deliberately (n = 43)1.00 1.00 
Self-screen in previous 3 yearsChi-square = 1.14, df = 1, P = .285Chi-square = 0.13, df = 1, P = 0.723
 Screened (n = 56)1.00 1.00 
 Not screened (n = 313)1.110.91-1.360.960.79-1.18
Skin reaction to summer sunChi-square = 0.12, df = 1, P = .728Chi-square = 0.02, df = 1, P = .892
 Always/usually burns (n = 191)1.00 1.00 
 Sometimes/never burns (n = 178)1.060.92-
Self-reported “moliness”Chi-square = 1.99, df = 1, P = .159Chi-square = 0.68, df = 1, P = .411
 Many/some (n = 153)1.00 1.00 
 Few/none (n = 215)1.110.96-

Because we were interested in understanding more about the factors related to the thickest and thinnest NM, we used multinomial regression models to assess the association between the same set of variables against tumor thickness (>4 mm, 1-4 mm, <1 mm) to produce odds ratios (ORs) and 95% confidence intervals (CIs). This latter analysis was adjusted for age and sex only because of the limited number of tumors in the 3 separate tumor groups.

For questions regarding clinical features and anatomic sites, we calculated percentages reporting each feature and compared responses for patients with thin (≤2.0 mm) NM, thick (>2.0 mm) NM, thin (≤2.0 mm) SSM, and thick (>2.0 mm) SSM. We specifically compared thin NM with thick NM, thin NM with thin SSM, and thick NM with thick SSM. For these analyses, we included only those melanomas detected by someone other than a physician. Significance across all 3 groups was assessed using chi-square statistic with 2 degrees of freedom, and 95% confidence intervals calculated for each percentage estimate.


Of the 369 respondents with NM, 60% were men. The median age of diagnosis was 57 years (range, 20-76 years), and 66% were 50 years of age or older. More than 56% had an educational level of grade 11 and above. About half (52%) reported that their skin always or usually burned upon first reaction to the summer sun. The median tumor thickness was 1.90 mm, ranging from 0.40 mm to 9.30 mm. Nodular melanoma was characterized in accordance with 2002 American Joint Committee on Cancer/International Union Against Cancer melanoma staging T classifications: ≤1 mm (T1, 22.2%), 1.01-2.0 mm (T2, 34.5%), 2.01-4.0 mm (T3, 30.6%), and >4.0 mm (T4, 12.7%).

Sixty-one percent of NM was self-detected, 27% detected by a partner, but only 11% detected by a physician. Only 53% of NMs were visible to the patient. Twenty-eight percent of patients had received a deliberate physician screen on all or nearly all of the whole body for early signs of skin cancer in the past 3 years, but only 15% of participants had performed a skin self-examination in the prior 3 years. Forty-two percent of patients reported that they had “many/some” moles.

For NM, men and older respondents were more likely to have tumors of greater thickness, after adjustment for all the other variables shown in Table 1. Thick NMs were also more likely in respondents who had not been screened by a physician in the past 3 years, and among those whose melanoma was on a visible part of their body. A nonsignificant trend toward thick NM was found among those who discovered their own lesion compared with a physician-detected NM. Seventy percent of visible lesions were first noticed by the individual, compared with 54% of nonvisible lesions, and partners were more likely to detect nonvisible lesions (31% vs 21%) (significant association between visibility and who first noticed the lesion, chi-square = 9.55, df = 2, P = .008).

Results from the multinomial model (adjusted for age and sex) demonstrated significant associations between NM thickness and age group, having had a clinical skin examination (CSE) in the last 3 years, whether the lesion was on a visible part of the body, who first noticed the lesion, and self-reported number of moles (Table 2). Specific analysis of NM of the extreme thickness categories (T1 and T4) revealed that respondents aged 60 to 75 years were nearly 3 times more likely to have a T4 NM than a T1 NM, compared with respondents aged 40 to 59 years (OR, 2.94; 95% CI, 1.33-6.47). Compared with respondents who had a CSE in the last 3 years, respondents who did not have a CSE were nearly 4 times more likely to have a T4 NM than a T1 NM (OR, 3.75; 95% CI, 1.47-9.59). In comparison with respondents whose NM was detected by a physician, those with self-detected melanoma were nearly 5 times more likely to have a T4 than a T1 (OR, 4.86; 95% CI, 1.21-19.51).

Table 2. Adjusted Odds Ratios of Thickest Nodular Melanoma (>4 mm) Compared With Thinnest Nodular Melanoma (≤1 mm) and Intermediate Nodular Melanoma (>1-4 mm) Compared With Thinnest Nodular Melanoma (≤1 mm)*
Variable1-4 mm vs <1 mm4 mm+ vs <1 mmOverall Significance
Estimate95% Confidence IntervalEstimate95% Confidence Interval
  • *

    Melanoma thickness was categorized into ≤1.00 mm, >1 mm to <4 mm, and >4 mm, and modeled using ordinal regression.

  • Estimates are the exponential of the relevant parameter estimate, and reflect the odds relative to the baseline of <1 mm.

  • All estimates are adjusted for age and sex.

Sex    Chi-square = 0.07, df = 2, P = .968
 Men (n = 222)0.910.54-1.550.810.36-1.79 
 Women (n = 147)1.00 1.00  
Age group    Chi-square = 14.85, df = 4, P = .005
 20-39 y (n = 52)1.050.52- 
 40-59 y (n = 164)1.00 1.00  
 60-75 y (n = 153)2.471.37-4.472.941.33-6.47 
Highest education    Chi-square = 2.69, df = 2, P = .261
 Up to year 10 (n = 163)1.320.76-2.291.650.74-3.71 
 Years 11-12 and upward (n = 205)1.00 1.00  
Physician screen in previous 3 years    Chi-square = 8.39, df = 2, P = .015
 Screened (n = 103)1.00 1.00  
 Not screened (n = 266)1.831.07-3.133.751.47-9.59 
Visibility of lesion    Chi-square = 7.51, df = 2, P = .023
 Not visible (n = 173)1.00 1.00  
 Visible (n = 196)1.570.93-2.672.861.25-6.54 
Who first noticed the melanoma?    Chi-square = 10.97, df = 4, P = .027
 Your partner (n = 99)2.260.95-5.371.500.31-7.18 
 Yourself (n = 228)1.550.71-3.374.861.21-19.51 
 Physician (n = 42)1.00 1.00  
Method of detection    Chi-square = 3.55, df = 2, P = .169
 By accident (n = 325)1.420.69-2.966.871.28-36.89 
 Deliberately (n = 43)1.00 1.00  
Self-screen in previous 3 years    Chi-square = 4.82, df = 2, P = .090
 Screened (n = 56)1.00 1.00  
 Not screened (n = 313)1.991.03-3.851.050.39-2.83 
Skin reaction to summer sun    Chi-square = 2.39, df = 2, P = .303
 Always/usually burns (n = 191)1.00 1.00  
 Sometimes/never burns (n = 178)1.220.72-2.050.770.34-1.74 
Self-reported “moliness”    Chi-square = 6.12, df = 2, P = .047
 Many/some (n = 153)1.00 1.00  
 Few/none (n = 215)1.851.10-3.111.320.61-2.86 

Clinical Features of Nodular Melanoma

Thin NM versus thick NM

Patients with thin NM were nearly twice as likely as patients with thick NM to report a change in color (46.2% vs 25.8%), 5 times more likely to observe an irregular shape, and far more likely to report brown pigmentation (47.4% vs 27.2%). Patients with thick NM more commonly reported a lumpy/raised appearance, bleeding/weeping, blister-like appearance, red/pink pigmentation, and paleness/no color. The scalp and neck were more commonly a site for thick NM, and the back of the trunk was more frequently a site of thin NM. However, for many other clinical features, such as change in size, irritation or onset of itch, or the lesion being “different from other spots” or having “just appeared”(1.4% of thin NM patients and 4.6% of thick NM patients)(Table 3), there were few differences between thin NM and thick NM.

Table 3. Percentage of Non–Physician-Detected Melanomas Exhibiting Selected Clinical Features and Anatomic Sites, by Thickness (<2 mm and ≥2.01 mm) and Histology (Nodular and SSM)
Feature*Thin (≤2.00 mm)Thick (2.01 mm+)
Nodular, % (n = 176)SSM, % (n = 1673)Nodular, % (n = 151)SSM, % (n = 81)
  • SSM indicates superficial spreading melanoma.

  • Percentages and 95% confidence intervals are weighted to account for undersampling of thinner melanomas.

  • *

    “On the [date first believed something was wrong] what was it about the spot that made you or someone else believe something might be wrong?”

  • Multiple responses possible.

  • Significance of overall differences between groups (chi-square test).

  • §

    P < .001.

  • P < .05.

  • “What color (or colors) was the spot?”

  • #

    P < .01.

Change in color46.2 (38.9-53.5)64.2 (62.3-66.2)25.8 (18.7-33.0)25.9 (16.1-35.7)§
Change in size22.1 (16.0-28.2)27.2 (25.4-29.1)21.2 (14.5-27.9)27.2 (17.2-37.1)
Irregular shape10.4 (6.0-14.9)12.9 (11.5-14.3)2.0 (0.0-4.3)4.9 (0.1-9.8)§
Irritation/itch15.9 (10.6-21.3)12.3 (11.0-13.7)13.9 (8.2-19.6)21.0 (11.9-30.1)
Lumpy/raised appearance31.5 (24.7-38.3)19.1 (17.5-20.7)42.4 (34.3-50.5)37.0 (26.2-47.9)§
“Different from other spots”9.2 (5.0-13.4)11.2 (9.9-12.5)9.3 (4.5-14.0)6.2 (0.8-11.6)
Bleeding/crusting/weeping17.3 (11.8-22.9)4.6 (3.7-5.4)31.8 (24.2-39.4)13.6 (5.9-21.3)§
“It just appeared”1.4 (0.0-3.1)3.1 (2.4-3.8)4.6 (1.2-8.1)1.2 (0.0-3.7)
Dry/scaly1.6 (0.0-3.4)2.6 (2.0-3.3)4.0 (0.8-7.2)3.7 (0.0-7.9)
Blister-like3.2 (0.6-5.8)1.4 (0.9-1.8)12.6 (7.2-18.0)4.9 (0.1-9.8)§
Pain2.5 (0.2-4.8)1.3 (0.8-1.8)4.0 (0.8-7.2)4.9 (0.1-9.8)
 Brown47.4 (40.1-54.7)63.1 (61.1-65.1)27.2 (19.9-34.4)37.0 (26.2-47.9)§
 Black28.5 (21.9-35.1)32.4 (30.5-34.4)21.9 (15.1-28.6)18.5 (9.8-27.2)#
 Red/pink19.7 (13.9-25.5)16.8 (15.3-18.3)39.1 (31.1-47.1)30.9 (20.5-41.2)§
 Pale/white/no color5.3 (2.0-8.5)3.8 (3.0-4.6)15.2 (9.4-21.1)11.1 (4.1-18.2)§
 Gray/blue3.2 (0.6-5.7)1.7 (1.2-2.3)3.3 (0.4-6.2)4.9 (0.1-9.8)
 Other12.3 (7.5-17.1)14.5 (13.1-16.0)15.9 (9.9-21.9)16.0 (7.8-24.3)
Body site    
 Scalp and neck11.6 (6.9-16.3)11.0 (9.7-12.3)23.2 (16.3-30.1)17.3 (8.8-25.8)§
 Upper extremities18.5 (12.8-24.2)20.4 (18.7-22.0)19.2 (12.8-25.7)23.5 (14.0-33.0)
 Lower extremities29.2 (22.6-35.9)28.2 (26.3-30.0)27.8 (20.5-35.1)18.5 (9.8-27.2)
 Back of trunk34.2 (27.2-41.1)32.0 (30.1-33.9)23.8 (16.9-30.8)29.6 (19.4-39.9)
 Front of trunk6.5 (2.9-10.1)8.4 (7.3-9.6)6.0 (2.1-9.8)11.1 (4.1-18.2)

Thin SSM versus thin NM

Patients with thin SSM were more likely than patients with a thin NM to report a change in color (64.2% vs 46.2%) and brown pigmentation (63.1% vs 47.4%). Bleeding/crusting/weeping and a lumpy/raised appearance were more common in thin NM than thin SSM. However, for many other clinical features, such as change in size, shape, “different from other spots,” lesion “just appearing,” and body site location, there were few differences between thin NM and thin SSM.

In general, there were few differences in clinical features between patients with thick NM and thick SSM in the proportion of respondents, although the former more commonly reported bleeding/crusting/weeping, blister-like appearance, and red/pink pigmentation.


In the United States, Australia, Italy, and elsewhere, the NM subtype accounts for a disproportionate fraction of thick cutaneous disease.10-11, 13 As melanoma thickness is strongly related to survival, gaining an understanding of this subtype is likely to be helpful in reducing melanoma mortality.

Four major findings emerge from this analysis of the Queensland population-based cancer registry. First, there have been few reports of the presentation of thin NM; herein, 57% of NMs were diagnosed ≤2.0 mm. This report coupled with a similar finding from the US SEER registry10 augurs well for identifying a significant subset of NM that can be detected relatively early. Second, well-established warning signs, known to be common for SSM, such as change in color and brown pigment, were also relatively common among patients diagnosed with thin NM. In fact, nearly half of all patients with thin NM reported a change in color. Third, social demographics known to be related to more deeply invasive melanoma, such as older age and male sex, were also associated with thick NM. Finally, the absence of a physician skin examination in the 3 years before diagnosis was also associated with thick NM.

Alerting the public and professionals alike to the warning signs of melanoma, while advocating early and prompt referrals for suspect lesions, has been the hallmark for education on early detection. Because of concerted early detection campaigns, progress in Australia and the US has been made in reducing the median tumor thickness of SSM, although the proportion of thick SSMs has remained virtually the same (Queensland Cancer Registry, personal communication).10 Less than 7% and 5% of all SSMs diagnosed in the US and Queensland, respectively, are >2 mm compared with similar rates in the 1980s. However, in both US and Queensland population-based cancer registries, NM is about ⅙ as common as SSM but 6 times more likely to result in thick melanoma (Queensland Cancer Registry, personal communication).10 Thus, enhancing earlier detection of nodular melanoma is likely to be important for reducing melanoma mortality.

Clinical Presentation of Nodular Melanoma

There have been few studies that have compared the clinical presentation of NM versus SSM or the clinical presentation of early versus late NM.12 Compared with SSMs, NMs have been commonly reported to be more symmetrical, of a single color, amelanotic, elevated, growing in diameter, and having a round border. Carli et al13 found few differences between NM and SSM in terms of anatomic site, presence of atypical nevi, and number of common nevi. In our study, warning signs generally associated with SSM (although not as common) were also present in many patients with thin NM: 47% had brown pigmentation, 46% had a change in color, and 22% had a change in size of their lesion. Very few patients with thin NM (1.4%) reported that the melanoma “just appeared.” Conversely, compared with SSM, patients with thick NM more commonly reported lumpy/raised appearance, paleness or no color, and a blister-like appearance, suggesting potential ulceration of the lesion. Likewise, lumpy/raised appearance was more common among thin NMs compared with SSMs.

In this study, 61% of NMs were self-detected, 27% were detected by a partner, but only 11% were detected by the physician. In contrast, most studies have demonstrated that about 25% of superficial spreading melanoma is first discovered by a physician, and that physician detection is uniformly associated with thinner disease.17-24 Herein, there was a trend toward more favorable NM being discovered by a physician (rather than self-detected), suggesting the need to emphasize active surveillance for change in color or size (among other factors) for warning signs that may be associated with early NM or SSM. Because the US Institute of Medicine recommendations25 alert physicians to suspicious signs in older persons, and because middle-aged and older patients who are most commonly diagnosed with NM11 also make frequent visits to their healthcare providers, physician-directed opportunistic screening might be suitable for detection of NM as well as for the more common SSM. Given that many patients with thin NM reported a change in color before diagnosis, the importance of careful and regular skin self-examination should be promoted as well.

Biologic Features of Nodular Melanoma

Notably, older respondents (aged 60-79 years) were 3 times more likely than persons aged 40 to 59 years to be diagnosed with thick NM and 12 times more likely than respondents aged 20 to 39 years. This finding could suggest that the biological features of melanoma may be different in middle-aged and older men and women, although thicker melanoma may also be attributed to other factors, such as the elderly's relative immunodeficiency, or that older individuals may have fewer social contacts and examine their own skin less often.

According to recent studies, the rate of growth of primary cutaneous melanoma appears to differ based on histologic subtype. Liu et al recently estimated that NM grows at 4 times the rate of either SSM or lentigo maligna melanoma, with an increase of nearly half a millimeter per month.26 Unlike the current study, where there was no apparent association between self-reported number of moles and NM thickness, Liu et al found an association between persons with few nevi and aggressive NM.26 Earlier, using a melanoma kinetic index, Grob et al found that aggressive tumor growth as reported by the patient rather than delay was responsible for thick melanoma.27 Therefore, it is likely that a previously undetermined subset of NMs is fast-growing without a detectable preclinical phase. Further epidemiologic and molecular studies should seek to determine ways of identifying persons at risk of faster-growing NM and explore the efficacy of more frequent and thorough screenings for those at greatest risk of fatal melanoma.

Our study assumes that NM is a distinct biologic entity rather than a tumorigenic (vertical) growth phase applicable to any subtype of melanoma in which the horizontal growth phase (nontumorigenic microinvasive or in situ component) is not evident. While controversy persists regarding the validity of subtyping melanoma, recent evidence of divergent pathways of melanomagenesis according to anatomic location, sun exposure, nevus count, age of onset, and various mutations (eg, B-RAF, c-Kit) suggest that distinct phenotypic and genotypic patterns of melanoma exist.28-32 Although not universally accepted, the current widespread use of World Health Organization (WHO) classification of melanoma lends validity to the recognition of NM as a distinct biologic entity.33 Congruent cancer registry data from Queensland and the SEER registry (Queensland Cancer Registry, personal communication)10 indicate that the proportion of all invasive melanomas that are of the nodular subtype is nearly equal in these 2 countries, which supports the concept that WHO classifications of melanoma subtype are similarly applied. As histologic subtype criteria undergo further refinement and reclassification, recognition of the nodular subtype accounting for disproportionate mortality and a significant fraction of all thick tumors is crucial to efforts to improve earlier diagnosis.


Although early NM was associated with a physician visit in the 3 years before diagnosis, there is no evidence from this study that patients exhibited any signs or symptoms of NM at that visit, nor is there confirmation that a physician screen at this visit found a lesion appearing to be an early NM. Moreover, the finding that thin NM was diagnosed in patients who had a physician screen in the 3 years before diagnosis may be unrelated to the physician examination and more a result of patients being aware of changing lesions.

Surprisingly, nonvisible NM was thinner than visible NM. Nonvisible lesions were more likely to be detected by the partner or physician, and partner or physician detection is more likely to be associated with thinner disease; however, a full explanation for thinner nonvisible NM remains unclear.

The coding of melanoma into the nodular and superficial spreading subtypes is completed internally by the Queensland Cancer Registry based on the information from state-wide pathology laboratories. Currently there has been no centralized histology review of these codes by independent experts. A preliminary check has indicated that a proportion of nodular melanomas arising within SSMs have been coded as nodular melanomas (rather than invasive SSMs), but we were unable to quantify this precisely. However, the finding that the proportion of nodular melanomas/all melanomas in Queensland has remained remarkably consistent for the past 20 years and that this rate is very close to those found in other population-based registries such as SEER should minimize concerns of differential reporting by pathology laboratories.


Our data highlight clinical and sociodemographic features (male sex and advanced age) associated with thicker NM. More importantly, this study revealed that thick NM was more common in individuals who did not have a physician skin examination in the previous 3 years and among those who discovered their own lesions, rather than having them detected by a physician. Well-established warning signs, heretofore associated with SSM (eg, change in color and brown pigmentation), were also relatively common among patients diagnosed with thin NM. Enhanced public and professional education to raise awareness of the existence and warning signs associated with thinner NM may ultimately assist in improving earlier detection and treatment of this subtype. Although we do not have definitive answers to the best methods for detecting earlier disease, this study raises the possibility that more frequent clinical examinations and skin self-examinations by trained, high-risk individuals can contribute to the reduction of thick, rapidly growing melanoma.

Conflict of Interest Disclosures

The authors made no disclosures.