Presentation and detection of invasive melanoma in a high-risk population

Presentation and detection of invasive melanoma in a high-risk population Michelle McPherson, MPH, a Mark Elwood, MD, b Dallas R. English, PhD, c Peter D. Baade, PhD, a Philippa H. Youl, MPH, a and Joanne F. Aitken, PhD a,d Brisbane, Queensland, Australia Background: Early detection of melanoma has been encouraged in Queensland for many years, yet little is known about the patterns of detection and the way in which they relate to tumor thickness. Objective: Our purpose was to describe current patterns of melanoma detection in Queensland. Methods: This was a population-based study, comprising 3772 Queensland residents diagnosed with a histologically confirmed melanoma between 2000 and 2003. Results: Almost half (44.0%) of the melanomas were detected by the patients themselves, with physicians detecting one fourth (25.3%) and partners one fifth (18.6%). Melanomas detected by doctors were more likely to be thin (\0.75 mm) than those detected by the patient or other layperson. Melanomas detected during a deliberate skin examination were thinner than those detected incidentally. Limitations: Although a participation rate of 78% was achieved, as in any survey, nonresponse bias cannot be completely excluded, and the ability of the results to be generalized to other geographical areas is unknown. Conclusion: There are clear differences in the depth distribution of melanoma in terms of method of detection and who detects the lesions that are consistent with, but do not automatically lead to, the conclusion that promoting active methods of detection may be beneficial. ( J Am Acad Dermatol 2006;54:783-92.) Queensland has the highest incidence of cutaneous melanoma in the world. 1,2 Recent estimates show that the lifetime risk of melanoma is 1 in 16 for men and 1 in 24 for women in Queensland. 3 This compares with US figures of 1 in 55 for men and 1 in 82 for women. 4 Mortality from melanoma has remained relatively stable in recent years, although the From Queensland Cancer Fund, Brisbane a ; National Cancer Control Initiative, Melbourne b ; Cancer Epidemiology Centre, The Cancer Council Victoria, Melbourne c ; School of Population Health, University of Queensland, Brisbane. d Funding source: National Health and Medical Research Council, Australia. Conflicts of interest: None identified. Accepted for publication August 4, 2005. Reprint requests: Dr Joanne Aitken, Queensland Cancer Fund, PO Box 201, Spring Hill, QLD 4005. E-mail: JAitken@qldcancer. com.au. 0190-9622/$32.00 ª 2006 by the American Academy of Dermatology, Inc. doi:10.1016/j.jaad.2005.08.065 incidence of melanoma in Queensland continues to increase at 2.6% and 1.2% per year for men and women, respectively. 5 Unlike other cancers, melanoma is usually visible and thus potentially detectable without the need for invasive procedures. Survival is dependent on thickness of melanoma 6 ; therefore detection of melanoma when it is in its thin, curable state affords the patient the best outcome. Internationally there is no evidence that the incidence of thick melanomas is decreasing. 7-10 Establishing who first notices melanomas, particularly those lesions that are diagnosed as thick, will help target future campaigns designed to reduce the incidence of thick, potentially fatal, melanoma. Previous studies have shown that between 40% and 74% of melanomas are detected by the patients themselves 11-19 and that women are more likely than men to detect melanoma. 11,14,15 Several studies have reported that doctors are more likely than other individuals to detect thin melanomas. 14-16,19-21 However, most of these studies have been limited by small samples, low response rates, 783

784 McPherson et al JAM ACAD DERMATOL MAY 2006 or nonrepresentative sampling methods that limit the extent to which the results can be generalized to the total population. This study is the largest population-based investigation of melanoma patients conducted in Australia. It will describe current patterns of melanoma detection in Queensland and determine which demographic, behavioral, and melanoma-specific characteristics are associated with melanoma detection. The results of this study will contribute to a better understanding of the process of melanoma detection and identify possible methods by which we can reduce the thickness at which melanomas are diagnosed. METHOD Case ascertainment Queensland residents between 20 and 75 years of age who were diagnosed with histologically confirmed first primary invasive cutaneous melanoma (not including acral lentiginous melanoma) between Jan 1, 2000 and Dec 31, 2003 were eligible for this study. During this period, 94.1% of all melanomas among Queensland residents were diagnosed within this age group. For sampling efficiency and cost, all patients with thick melanomas ($0.75 mm) and a random 60% sample of those with thinner melanomas (\0.75 mm) were selected. Patients with noncutaneous melanoma, metastatic disease with unknown primary lesions, or a confirmed previous melanoma were excluded. A few patients were simultaneously diagnosed with more than one melanoma; for these patients the thickest melanoma was selected for this study. Eligible patients were identified from the Queensland Cancer Registry, to which cancer notifications have been a legal requirement since 1982. Letters explaining the study and seeking permission to contact the patient were sent to all treating doctors. Nonresponding doctors were telephoned after 1 week After permission was obtained, patients were invited by letter to participate. Ethics Ethical clearance for this study was obtained from the University of Queensland Behavioural and Social Science Ethical Review Committee. Data collection Data were collected from histopathology forms and from the patients by trained interviewers using computer-assisted telephone interview. The structured interview collected information on demographics; sun exposure and sunburn history; detection, diagnosis, and treatment of the melanoma; history of melanoma and other cancers; as well as skinexamination behaviors. The median time between diagnosis and interview was 5 months (mean, 5.5 months; range, 1-26 months). Almost three fourths of patients (71.3%) were interviewed within 6 months of their diagnosis. To ascertain how melanoma was detected, patients were asked, Who was the very first person who first believed that something might be wrong with the spot that was the melanoma? Possible responses included yourself, your partner, other relative, a friend, other lay people (someone who was not a doctor), or a doctor. Patients were also asked to indicate whether the melanoma had been detected incidentally or during a deliberate skin examination. Symptoms were ascertained by asking, in an open-ended fashion, What was it about the spot that made you or someone else believe something might be wrong? The interviewer then selected the most appropriate category for the response. These categories included a change in size, color, or shape; a lumpy/raised appearance; bleeding/crusting; irritation/itch; pain; and it just looked different from other spots. Patients were additionally asked an open-ended question: What color or colors were the spot? Because patients would not be able to report on signs noticed by a doctor, symptom information was available for non-doctor-detected melanoma only. Clinical skin examination was determined by asking: During the last 3 years before you first believed something was wrong, had a doctor deliberately checked all or nearly all of your whole body for the early signs of skin cancer? The validity of this question was recently quantified in a separate study by comparing responses against medical records. 22 Concordance between self-reported clinical skin examinations and medical records was high at 93%, as was sensitivity (92%) and specificity (96%). Similarly, history of skin self-examination was determined by asking: During the last 3 years before you first believed something was wrong, had you or someone else who was not a doctor deliberately checked all or nearly all of your whole body for early signs of skin cancer? Patients were asked during the interview to indicate (based on a diagram of the body mailed before the interview) where on the body their melanoma was. This information was then used to determine visibility categories. In all cases, the reported site was checked against the site recorded on the pathology report and discrepancies were minimal, none that altered the visibility category. Lesions were determined to be nonvisible if they were on the scalp,

JAM ACAD DERMATOL VOLUME 54, NUMBER 5 McPherson et al 785 back of the neck, ears and shoulders, back, buttocks, back of the legs, and soles. The remaining melanomas were considered to be on visible sites of the body. This was consistent with other studies that reported on melanoma by visibility. 16,23,24 Socioeconomic status was determined by using an index of disadvantage derived from geographical area of residence. 25 Remoteness, based on geographical area of residence, was assessed by using the Accessibility/Remoteness Index of Australia classification, 26 with the categories dichotomized to highly accessible versus the remainder. Test-retest reliability The test-retest reliability of the questionnaire was assessed by reinterviewing 186 patients 1 to 3 months after their first interview. There was good complete agreement for the main outcome measure ( Who was the first person who first believed that something might be wrong with the spot that was the melanoma? ) as 83.8% reported the same in both interviews. For the symptoms recorded, this agreement ranged from 76.6% for change in colour to 94.2% for bleeding/weeping. Agreement was also high for method of detection (85.9%), clinical skin examination (86.8%), and skin self-examination (92.9%) in the past 3 years. Statistical analysis The proportion of thin melanomas was weighted by 1.67 to account for the random selection of 60% of the thin melanomas from the Queensland Cancer Registry. Except for the sample description and multivariate analyses, all reported percentages reflect this weighting. Statistical testing of bivariate comparisons was performed by means of the x 2 test. Factors that were statistically significant (P \.05) were then included in a multivariate model. Analysis of factors that were associated with who first noticed the melanoma was performed by logistic regression analysis. A separate logistic regression model was used to investigate reported signs and symptoms and their relation to thickness. Indicator variables for each sign and symptom were included in the model, along with potential confounders of sex and age. Thickness (the outcome variable) was grouped into 3 categories (\0.75, 0.75-2.99, $3 mm). All data were analyzed by means of SAS software. 27 RESULTS Response We identified 4854 eligible patients, but 15 were deceased and doctor s consent had not been obtained for 329. Of the remaining 4510 patients, 3887 consented to participate, 369 refused, 201 were contacted but never replied, and 53 were unable to be contacted (no current address or telephone number). Of the total, 80% of patients agreed to participate and 3772 (78%) completed an interview. There were no significant differences between participants and nonparticipants in relation to age, site, morphology of melanoma, or geographical location and socioeconomic status of the person s place of usual residence. Nonparticipants were significantly more likely to be men (63.4% compared with 57.1%; P \.01) and to have a thick melanoma (10.9% compared with 7.2%; P \.01). The median thickness for participants was 0.68 mm and for nonparticipants, 0.75 mm. The final sample included 2150 men (57.0%) and 1622 women (43.0%). The median age at diagnosis was 55 years for men and 50 years for women. In accordance with our sampling scheme, approximately half (54.3%) of the sample were diagnosed with a thin melanoma (\0.75 mm). All percentages in the following sections have been weighted to reflect the actual thickness distribution of melanomas in the Queensland population to account for our deliberate undersampling of thin melanomas. Who first notices melanoma? Nearly half of the patients reported first noticing the melanoma themselves (44.0%), whereas one fourth (25.3%) of melanomas were first noticed by a doctor (Table I). Partners first noticed 18.6% of melanomas, and other laypersons (including other relatives, friends, and service people [eg, hairdressers]) detected 12.1%. Compared with men, a greater proportion of women detected their melanomas themselves (57.1% vs 33.8%), whereas men had a greater proportion of partner-detected melanoma (26.7% vs 8.1%) (Table I). General practitioners were responsible for the majority of doctor-detected melanomas (54.1%). Other melanomas were detected by a doctor in a skin clinic (21.5%) or a dermatologist (17.4%), with a smaller proportion detected by a surgeon, public hospital doctor, or other specialist (7.0% combined). Doctors working in skin clinics and dermatologists detected melanomas almost exclusively during scheduled skin examinations (99.0% and 96.6%, respectively). General practitioners detected equally during a deliberate examination and incidentally (54.6% compared with 44.9%), as did surgeons (44.2% compared with 55.8%). For other specialists, only 13.3% of melanomas were detected during a deliberate skin examination.

786 McPherson et al JAM ACAD DERMATOL MAY 2006 Table I. Who first notices melanoma, according to characteristics of the patient and the melanoma* Characteristic Doctor (n = 881) Patient (n = 1732) Partner (n = 707) Other layperson (n = 452) P value y All (N = 3772) 25.3 44.0 18.6 12.1 Demographics Sex Male (n = 2150) 30.0 33.8 26.7 9.5 Female (n = 1622) 19.2 57.1 8.1 15.6 Age (y) 20-49 (n = 1490) 17.0 49.5 18.4 15.1 50-76 30.9 40.2 18.7 10.2 SES group.007 Affluent (n = 263) 28.7 50.4 15.2 5.7 Middle (n = 3359) 25.0 43.4 19.1 12.5 Disadvantaged (n = 150) 24.1 45.0 14.4 16.5 Marital status z Married (n = 2897) 25.3 42.1 23.3 9.3 Not married (n = 869) 25.2 50.0 3.1 21.7 Melanoma characteristics Morphology.601 SSM (n = 2327) 25.9 42.3 18.9 12.9 NM (n = 369) 11.8 61.5 17.3 9.4 LLM (n = 128) 36.4 33.3 16.7 13.6 Other (n = 126) 22.4 52.1 17.8 7.7 NOS (n = 822) 26.6 43.5 18.5 11.4 Visibility Visible (n = 1711) 19.0 58.3 12.3 10.4 Not visible (n = 2061) 30.4 32.3 23.7 13.6 Site Head and neck (n = 508) 17.9 45.3 16.9 19.9 Arms (n = 711) 22.7 55.7 10.8 10.8 Legs (n = 871) 12.5 65.3 11.5 10.7 Front of trunk (n = 315) 28.1 50.3 16.6 5.0 Back of trunk (n = 1367) 36.2 23.1 28.0 12.7 Thickness (mm) \0.75 (n = 2049) 30.5 38.6 18.2 12.7 0.75-1.49 (n = 1017) 17.6 49.8 20.6 12.0 1.5-2.9 (n = 443) 10.8 58.7 18.1 12.4 [3.0 (n = 253) 9.9 67.6 17.0 5.5 Detection and screening Detection z Deliberate skin check (n = 870) 70.8 18.4 9.3 1.5 Incidentally (n = 2895) 10.0 52.5 21.8 15.7 CSE in past 3 y Yes (n = 1284) 46.1 35.2 12.9 5.8 No (n = 2488) 13.9 48.7 21/7 15.6 SSE in past 3 y Yes (n = 640) 20.4 49.4 22.5 7.6 No (n = 3132) 26.3 42.8 17.8 13.1 CSE, Clinical skin examination; LLM, lentigo maligna melanoma; NM, nodular melanoma; NOS, not otherwise specified; SES, socioeconomic status; SSE, skin self-examination; SSM, superficial spreading melanoma. *Proportions have been weighted to account for 60% sampling of thin (\0.75 mm) melanomas. Data given as percentages. y Chi-square test. z Total is less than 3772 because of missing responses. Includes desmoplastic, spindle cell, amelanotic melanoma, and melanoma in a junctional nevus.

JAM ACAD DERMATOL VOLUME 54, NUMBER 5 McPherson et al 787 Overall, approximately half (45.1%) of the sample had a melanoma arising on a site that was easily visible. Visible melanoma was more commonly detected by the patient, whereas nonvisible melanoma was detected equally by patients and doctors. Compared with other sites, melanoma occurring on the back was more commonly detected by a doctor or partner (Table I). More than one third (35.3%) of patients reported having a whole-body clinical skin examination in the 3 years before their melanoma was detected, and 17.2% said they had performed a whole-body skin self-examination during that period. Almost half (46.1%) of those patients reporting a clinical skin examination in the past 3 years indicated that melanoma had been detected by a doctor, compared with 13.9% who did not report having a clinical skin examination. To examine predictors of melanoma detection according to who detected the melanoma, we constructed a multivariate model comparing layperson-detected with doctor-detected melanoma. All variables that were statistically significant (P\.05) at the bivariate level were included in a multivariate logistic regression model (Table II). Compared with patients whose melanomas were detected by a doctor, patients were more likely to have had their melanoma detected by a layperson (including themselves) if they were female or younger, if the melanoma was thicker than 0.75 mm at diagnosis, and if it was on a visible site. Patients were also more likely to have had their melanoma detected incidentally and to have performed a skin self-examination in the 3 years before their melanoma was detected. They were less likely to have reported a clinical skin examination in the past 3 years (Table II). Melanoma thickness and methods of detection The thickness distribution of melanomas varied markedly according to the person who first detected the melanoma (doctor or layperson). Overall, 80.6% of lesions detected by a doctor were thin (\0.75 mm) compared with 61.9% of lesions detected by laypersons (Table III). Within each group of patients, melanomas detected during a deliberate skin examination had a more favorable depth distribution than melanomas detected incidentally (84.6% vs 71.2% of thin lesions for doctor-detected melanomas and 71.4% vs 60.8% for lesions detected by the patients themselves or another layperson) (Table III). We performed multivariate analysis separately for patients whose melanomas were detected by doctors and laypersons, adjusting for factors that were Table II. Associations between patients and melanoma characteristics and likelihood of doctors detecting the melanoma: Results of multivariate logistic regression model Characteristic Doctor detected, no. * Odds ratio (95% CI) y Non-doctor-detected: Doctor-detected z P value Sex.003 Male 595 (30.0) 1.0 Female 286 (19.2) 1.38 (1.11-1.70) Age (y) 20-49 234 (17.0) 1.0 50-76 647 (30.9) 0.45 (0.36-0.56) ARIA category.534 Highly accessible 674 (26.8) 1.0 Remainder 207 (21.4) 0.93 (0.74-1.17) Morphology.152 SSM 573 (25.9) 1.0 NM 42 (11.8) 1.29 (0.82-2.04) LLM 45 (36.4) 0.59 (0.36-0.98) NOS 197 (22.4) 0.93 (0.75-1.20) Other k 24 (26.6) 0.77 (0.41-1.42) Visibility Not visible 306 (19.0) 1.0 Visible 575 (30.4) 1.59 (1.29-1.96) Thickness (mm) \0.75 625 (30.5) 1.0 0.75-1.49 179 (17.6) 1.56 (1.22-1.99) 1.5-2.9 48 (10.8) 2.75 (1.85-4.11) $3.0 25 (9.9) 2.26 (1.31-3.90) Detection Deliberate skin check 602 (70.8) 1.0 Incidentally 275 (10.0) 19.63 (15.49-24.88) CSE in past 3 y No 557 (13.9) 1.0 Yes 324 (46.1) 0.42 (0.34-0.53) SSE in past 3 y No 120 (26.3) 1.0 Yes 761 (20.4) 4.53 (3.38-6.08) ARIA, Accessibility/Remoteness Index of Australia; CI, confidence interval; CSE, clinical skin examination; LLM, lentigo maligna melanoma; NM, nodular melanoma; NOS, not otherwise specified; SSE, skin self-examination; SSM, superficial spreading melanoma. *Odds ratios and 95% confidence intervals are adjusted for all other variables in the table. y Proportions have been weighted to account for 60% sampling of thin (\0.76 mm) melanomas. z Non-doctor-detected included patient-detected, partnerdetected, and layperson-detected melanomas. P value based on the chi-square statistic from the logistic model representing the significance of the specific variable. k Includes desmoplastic, spindle cell, amelanotic melanoma, and melanoma in a junctional nevus. significant at the bivariate level, including sex, age, education level, as well as thickness and visibility of melanoma. In both groups of patients, melanoma detected during a deliberate skin check was more

788 McPherson et al JAM ACAD DERMATOL MAY 2006 Table III. Detection methods and screening practices by the person who detected the melanoma and by thickness* Melanoma detected by doctor Melanoma detected by patient/partner/other Characteristic (N = 3762) Total no. \0.75 y 0.75-1.49 1.50-2.99 [3.0 Total no. \0.75 0.75-1.49 1.50-2.99 $3.0 Total 877 80.6 13.8 3.7 1.9 2885 61.9 21.8 10.3 5.9 Detection z During deliberate 600 84.6 11.3 3.0 1.1 267 71.4 18.7 7.8 2.1 skin check Incidental 273 71.2 19.9 5.2 3.7 2615 60.8 22.2 10.6 6.4 Screening in past 3 y CSE only 458 83.0 13.0 2.5 1.5 486 62.0 22.3 11.1 4.6 Both CSE and SSE 23 85.9 7.4 4.7 2.0 279 67.8 19.4 8.5 4.3 SSE only 97 75.8 15.1 6.1 3.0 240 63.0 22.0 8.8 6.2 None 299 75.0 17.3 5.1 2.6 1880 61.0 22.0 10.5 6.5 CSE, Clinical skin examination; SSE, skin self-examination. *Proportions have been weighted to account for 60% sampling of thin (\0.75 mm) melanomas and patients with unknown melanoma thickness (n = 10) have been excluded from the analysis. y Thickness measured in millimeters. z Totals do not necessarily add up to 3762 because of missing responses. likely to be thin than melanoma detected incidentally. However, this association was slightly stronger when the melanoma was detected by a doctor (odds ratio 2.57; 95% confidence level [CI], 1.84-3.59 compared with odds ratio 1.59; 95% CI, 1.20-2.10). The associations between these two factors (ie, doctor vs layperson and deliberate skin check vs incidental discovery) and melanoma thickness were largely independent. In an analysis of all melanomas, regardless of who detected them, the odds of detecting a thin melanoma were 2.56 times greater if the melanoma had been detected by a doctor rather than by a layperson (95% CI, 2.17-3.01). After adjusting for method of detection (deliberate vs incidental), the odds of detecting a thin melanoma were still 1.81 times greater (95% CI 1.49-2.21) if the melanoma had been detected by a doctor. Overall, more than two thirds of all melanomas (68%) were detected by the patient or another layperson as an incidental finding, and this group had the least favorable thickness distribution, with 61% of these being thinner than 0.75 mm and 6.4% being 3 mm and thicker at diagnosis. In comparison, lesions detected during a deliberate skin examination by a doctor have the most favorable thickness distribution, with 85% being thinner than 0.75 mm thick and only 1% being thicker than 3 mm. Melanomas detected by a doctor tended to be thinner when patients reported a clinical skin examination in the 3 years before their melanoma was detected. Among melanomas detected by laypersons, only minor variations in thickness distribution were found according to reported previous screening behavior (either clinical skin examination or skin self-examination) (Table III). Signs and symptoms Most patients (65.2%) reported already having a spot at the place where the melanoma started, 12.0% indicated they did not have a spot, and 22.8% did not know. Of the 2463 patients who reported having a spot at the place where the melanoma started, 28.4% said the spot had been there all their life, 26.5% for years, 15.2% for months, and 1.9% for weeks or days. More than one fourth (28.0%) did not know how long they had had the spot. Patients whose melanoma was not detected by a doctor (n = 2885) most commonly reported signs and symptoms such as changes in color, size, or shape, or an irritation or itch; they also reported that the lesion had a lumpy or raised appearance and that it looked different from other spots (Table IV). Approximately one third of patients (37.2%) reported only one sign or symptom, 39.7% reported two, 18.3% reported 3, and 4.8% reported more than 3. When more than one symptom was reported, the number of different combinations was too great for meaningful analysis. The most commonly reported pigmentation was brown, followed by black and red/pink (Table IV). The most common signs or symptoms reported by patients diagnosed with thin melanomas were a change in color or size and brown or black pigmentation (Table IV). The most common signs or symptoms reported by patients with the thickest melanomas were change in shape, bleeding/weeping, and a change in color or size; the most common pigmentation was red/pink (Table IV).

JAM ACAD DERMATOL VOLUME 54, NUMBER 5 McPherson et al 789 DISCUSSION The vast majority of melanomas were first detected by the patients themselves, their partners, and laypersons. However, melanomas detected by physicians were more likely to be thin at diagnosis than those detected by laypersons. Similarly, lesions detected by a doctor at a routine skin examination had a much more favorable thickness distribution than those detected by the patients themselves or another layperson as an incidental finding. To our knowledge, this is the largest populationbased study, and the first such study in Australia, to assess who first notices melanoma and the methods of melanoma detection. Our very high response rate and lack of nonresponse bias suggest that these results have high validity. Because data for this study were based mainly on self-reports, there is a potential for recall bias. However, most participants were interviewed within 6 months of their diagnosis. Data from our reliability study suggest that the measures used in this study had high test-retest reliability. Detection by a layperson was more common for female patients, young persons, and those with melanoma on a visible site. The higher proportion of self-detected melanomas among women and partner-detected melanomas among men highlight the substantial contribution that women make in the detection of melanoma. Consistent with this finding, other studies have suggested that this sex difference was due to women being more conscious of appearance than men and having greater awareness of their own skin, 11,14,15 even though they may not necessarily conduct more self-examinations of their skin (as found in this study and by Aitken et al 28 ). Women s greater knowledge of melanoma 29-31 and higher prevalence of preventive behaviors 32 may also contribute to the difference. Our finding that older patients ($50 years of age) were more likely to have their melanoma detected by a doctor is consistent with those findings reported elsewhere. 11,16 Possible reasons for this might include more frequent visits to a doctor, both for specific problems and for routine health checks 33 and being more likely to have a whole-body skin examination by a doctor. 34 This detection differential may also be due to a reduced ability and confidence among older people to recognize change in a skin lesion themselves. 35 Overall, melanoma in Queensland is usually diagnosed when it is thin. Between 2000 and 2002 in Queensland, among melanomas with known thickness (94% of the total), 63% of lesions were thinner than 0.75 mm at diagnosis, 29% were Table IV. Reported symptoms for patient-/ layperson-detected melanoma by thickness* Signs/symptoms (N = 2885) Total \0.75 (n = 1424) Thickness (mm) 0.75-2.99 (n = 1233) $3.00 (n = 228) P value y Change in color 58.1 65.6 49.9 24.6 Change in size 25.6 27.9 21.4 23.7.030 Change in shape 21.6 15.0 30.3 43.0 Irritation/itch 13.2 10.8 17.4 15.4.028 Lumpy/raised 11.8 14.1 8.9 2.6.009 appearance Different from 10.4 11.3 9.2 7.5.102 other spots Bleeding/weeping 7.2 3.0 11.9 25.0 It just appeared 2.9 2.9 3.0 2.2.719 Dry/scaly 2.7 2.4 3.6 2.3.147 Blister-like 2.5 1.1 3.5 11.4 Pain 1.8 1.1 2.4 5.3.018 Pigmentation Brown 58.7 65.0 53.0 23.2.631 Black 30.0 31.7 28.3 21.5 Red/pink 18.5 15.9 20.2 37.3.050 Pale/no color 5.3 3.2 6.9 17.5.002 Purple 2.7 2.1 3.3 5.3.322 Gray/blue 2.2 2.0 2.7 1.8.480 *Proportions weighted to account for 60% sampling of thin (\0.75 mm) melanomas. Multiple responses allowed. Patients with unknown melanoma thickness (n = 6) have been excluded from the analysis. y Multivariate analysis adjusted by age and sex. between 0.75 and 2.99 mm, and 8% were 3 mm or thicker (Queensland Cancer Registry, written communication, June 2005). Our results indicate that this distribution is likely to vary by whether the lesion is first detected by a doctor or by another person, including the patient, and whether it is detected at a routine skin examination or incidentally. Doctors were more likely than laypersons to detect melanomas when thin; for both doctor- and laypersondetected melanomas, the lesion was more likely to be thin if it had been detected as part of a routine skin examination than incidentally. The more favorable thickness distribution associated with routine skin examinations than with incidental findings is to be expected because presumably the less obvious visual characteristics that can be found during a routine skin examination are present earlier than the more obvious signs or symptoms that would trigger action on the basis of an incidental finding. This applies whether doctors or laypersons are conducting the examinations. In general, the more favorable thickness distribution of lesions detected by doctors is presumably due to more thorough checking of the skin done by doctors

790 McPherson et al JAM ACAD DERMATOL MAY 2006 as well as the greater awareness by doctor of skin abnormalities. Similar findings have been reported by several other international studies. 15,16,18,20 The most commonly reported signs and symptoms in this study were similar to those reported elsewhere 13,19,21,36 and included changes in color, size, and shape; an irritation or itch; and the appearance of the lesion as lumpy or raised. However, one striking result from this study is the increased proportion of persons diagnosed with thick melanoma reporting red/pink pigmentation, pale or no color, and the appearance of the melanoma as a blister. These latter symptoms or signs are not part of the commonly used symptom checklists, including the ABCD system 37 or MacKie s 7-point checklist 38 ; therefore these symptoms may not be a warning to the patient that the lesions may be melanoma. These checklists may need to be revised to include the fact that these characteristics may be suggestive of melanoma. Possible methods to improve early detection These data imply that if a greater proportion of people received skin examinations from their doctor or if routine skin examinations were carried out more frequently by patients themselves or by other laypersons, the thickness distribution of diagnosed lesions might improve. However, currently there is no definitive evidence from randomized trials to suggest that population skin screening would be effective in reducing mortality associated with malignant lesions 39,40 ; therefore reputable groups applying evidence-based methods have recommended against routine skin examination either by doctors or by the public. 41,42 Providing support and funding for routine skin screening by doctors as a public health measure is a major decision with significant consequences in terms of possibly unnecessary diagnoses and excisions as well as costs related to the screenings. The lack of objective evidence of effectiveness in terms of a reduction in mortality associated with melanoma or in the population-based incidence of deeply invasive melanoma is a major barrier to implementation. Another method to improve early detection would be to improve the ability of laypersons to identify suspect skin lesions at an earlier stage. However, differentiating early melanomas from benign lesions is difficult for the public. A study using computer-generated changes in lesions showed that people are unable to memorize their skin sufficiently well to recognize changes over long intervals. 43 This is reinforced by the observation that, compared with doctor-detected lesions, lesions detected by laypersons were 20 times more likely to be noticed incidentally than as a result of a systematic skin check. Nevertheless, there may be benefit in encouraging people to become familiar with their skin and to present any changes to a physician. This mirrors the current advice given for early detection of breast cancer as women are asked to be aware of how their breasts normally appear and feel and to report any new or unusual changes promptly to their general practitioner. 44,45 Increasing the practice of skin examinations either by doctors or by laypersons, encouraging familiarity of the skin among laypersons with presentation of suspect lesions to a doctor, and modifying published symptom checklists may shift the thickness distribution of melanoma toward a higher proportion of thin melanomas, although with the potentially greater cost associated with increased excision of benign lesions, worry, and financial cost, none of which were assessed in this study. Even more importantly, it has not been shown that the increased detection of thin melanoma corresponds to a reduction in the incidence of thick melanoma and an improvement in survival. Instead this may reflect increased diagnosis of very slowly progressive lesions that may have little effect on mortality associated with melanoma. 46-48 Consistent with this hypothesis, the increase in detected thin melanomas in recent decades has not been accompanied by an equivalent decrease in thick melanomas. 49 CONCLUSIONS The message of this observational study is that there are clear differences in the thickness distribution of melanoma in terms of method of detection and who detects the lesions, and perhaps in symptomatology, which are consistent with, but do not automatically lead to, the conclusion that promoting active methods of detection may be beneficial. This study emphasizes the potential for earlier diagnosis and the need for definitive studies to address this issue. REFERENCES 1. Australian Institute of Health and Welfare Australasian Association of Cancer Registries. Cancer in Australia 2001. Canberra: Australian Institute of Health and Welfare; 2003. 2. Parkin DM, Muir C, Whelan SL. Cancer incidence in five continents. New York: Oxford University Press; 2002. 3. Baade P, Coory M, Ring I. National Health Priority cancers in Queensland (1982-1997). Available at: http://www.health.qld. gov.au/publications/nhpa.pdf. Accessed April 22, 2005. 4. Jemal A, Tiwari RC, Murray T, Ghafoor A, Samuels A, Ward E, et al. Cancer statistics, 2004. CA Cancer J Clin 2004;54:8-29. 5. Hall L, Youden D, Coory M. Mortality and incidence trends for leading cancers in Queensland, 1982 to 2002. Available at:

JAM ACAD DERMATOL VOLUME 54, NUMBER 5 McPherson et al 791 http://www.health.qld.gov.au/publications/infocirc/info68.pdf. Accessed April 22, 2005. 6. South Australian Cancer Registry. Epidemiology of cancer in South Australia. Incidence, mortality and survival 1977 to 1997. Adelaide: South Australian Cancer Registry; 1998. 7. Jemal A, Devesa SS, Hartge P, Tucker MA. Recent trends in cutaneous melanoma incidence among whites in the United States. J Natl Cancer Inst 2001;93:678-83. 8. Lipsker D, Hedelin G, Heid E, Grosshans E, Cribier B. Striking increase of thin melanomas contrasts with stable incidence of thick melanomas. Arch Dermatol 1999;135:1451-6. 9. Balzi D, Carli P, Giannotti B, Paci E, Buiatti E. Cutaneous melanoma in the Florentine area, Italy: incidence, survival and mortality between 1985 and 1994. Eur J Cancer Prev 2003;12:43-8. 10. Marrett LD, Nguyen HL, Armstrong BK. Trends in the incidence of cutaneous malignant melanoma in New South Wales, 1983-1996. Int J Cancer 2001;92:457-62. 11. Richard MA, Grob JJ, Avril MF, Delaunay M, Gouvernet J, Wolkenstein P, et al. Delays in diagnosis and melanoma prognosis (I): the role of patients. Int J Cancer 2000; 89:271-9. 12. Schmid-Wendtner MH, Baumert J, Stange J, Volkenandt M. Delay in the diagnosis of cutaneous melanoma: an analysis of 233 patients. Melanoma Res 2002;12:389-94. 13. Blum A, Brand CU, Ellwanger U, Schlangerhauff B, Stroebel W, Rassner G, et al. Awareness and early detection of cutaneous melanoma: an analysis of factors related to delay in treatment. Br J Dermatol 1999;141:783-7. 14. Koh HK, Miller DR, Geller AC, Clapp RW, Mercer MB, Lew RA. Who discovers melanoma? Patterns from a population-based survey. J Am Acad Dermatol 1992;26:914-9. 15. Epstein D, Lange J, Gruber S, Mofid M, Koch S. Is physician detection associated with thinner melanomas? JAMA 1999; 281:640-2. 16. Brady M, Oliveria S, Christos P, Berwick M, Coit D, Katz J, et al. Patterns of detection in patients with cutaneous melanoma. Cancer 2000;89:342-7. 17. Brochez L, Verhaeghe E, Bleyen L, Naeyaert J. Time delays and related factors in the diagnosis of cutaneous melanoma. Eur J Cancer 2001;37:843-8. 18. Carli P, De Giorgi V, Palli D, Maurichi A, Mulas P, Orlandi C, et al. Dermatologist detection and skin self-examination are associated with thinner melanomas: results from a survey of the Italian Multidisciplinary Group on Melanoma. Arch Dermatol 2003;139:607-12. 19. Elwood JM, Gallagher R. The first signs and symptoms of melanoma: a population-based study. Pigment Cell 1988;9: 118-30. 20. Richard MA, Grob JJ, Avril MF, Delaunay M, Gouvernet J, Wolkenstein P, et al. Delays in diagnosis and melanoma prognosis (II): the role of doctors. Int J Cancer 2000;89:280-5. 21. Krige JE, Isaacs S, Hudson DA, King HS, Strover RM, Johnson CA. Delay in the diagnosis of cutaneous malignant melanoma. Cancer 1991;68:2064-8. 22. Aitken JF, Youl PH, Janda M, Elwood M, Ring IT, Lowe JB, et al. Validity of self-reported skin screening histories. Am J Epidemiol 2004;159:1098-105. 23. Nagore E, Oliver V, Moreno-Pico S, Fortea J. Primary cutaneous melanoma in hidden sites is associated with thicker tumours a study of 829 patients. Eur J Cancer 2001;37:79-82. 24. Montella M, Crispo A, Grimaldi M, De Marco MR, Ascierto P, Parasole R, et al. An assessment of factors related to tumor thickness and delay in diagnosis of melanoma in southern Italy. Prev Med 2002;35:271-7. 25. ABS. Information paper: socio-economic indexes for areas, Australia, 2001. Canberra: Australian Bureau of Statistics; 2003. 26. Australian Institute of Health and Welfare. Rural, regional and remote health: a guide to remoteness classifications. Available at: http://www.aihw.gov.au/publications/index.cfm/title/9993. Accessed April 22, 2005. 27. SAS OnlineDocÒ, Version 9.1.2. Cary (NC): SAS Institute; 2004. 28. Aitken J, Janda M, Lowe JB, Elwood JM, Ring I, Youl P, et al. Prevalence of whole-body skin self-examination in a population at high risk for skin cancer (Australia). Cancer Causes Control 2004;15:453-63. 29. Melia J, Ellman R, Chamberlain J. Investigating changes in awareness about cutaneous malignant melanoma in Britain using the Omnibus survey. Clin Exp Dermatol 1994; 19:375-9. 30. Miller DR, Geller AC, Wyatt SW, Halpern A, Howell JB, Cockerell C, et al. Melanoma awareness and self-examination practices: results of a United States survey. J Am Acad Dermatol 1996; 34:962-70. 31. Brandberg Y, Bolund C, Michelson H, Mansson-Brahme E, Ringborg U, Sjoden PO. Perceived susceptibility to and knowledge of malignant melanoma: screening participants vs the general population. Prev Med 1996;25:170-7. 32. Mechanic D, Cleary P. Factors associated with the maintenance of positive health behavior. Prev Med 1980;9:805-14. 33. Heywood A, Sanson-Fisher RW, Ring I, Mudge P. Risk prevalence and screening for cancer by general practitioners. Prev Med 1994;23:152-9. 34. Janda M, Elwood M, Ring I, Firman D, Lowe J, Youl P, et al. Prevalence of skin screening by general practitioners in regional Queensland. Med J Aust 2004;180:10-5. 35. Hanrahan P, Hersey P, D Este C. Factors involved in presentation of older persons with thick melanomas. Med J Aust 1998;169:410-4. 36. Cassileth B, Temoshok L, Frederick BE, Walsh WP, Hurwitz S, Guerry D, et al. Patient and physician delay in melanoma diagnosis. J Am Acad Dermatol 1988;18:591-8. 37. Friedman RJ, Rigel DS, Silverman MK, Kopf AW, Vossaert KA. Malignant melanoma in the 1990 s. The continued importance of early detection and the role of physician examination and self-examination of the skin. CA Cancer J Clin 1991;41:201-26. 38. MacKie R. Clinical recognition of early invasive malignant melanoma. BMJ 1990;301:1005-6. 39. Aitken JF, Elwood JM, Lowe JB, Firman DW, Balanda KP, Ring I. A randomized trial of population screening for melanoma. J Med Screen 2002;9:33-7. 40. Helfand M, Mahon SM, Eden KB, Frame PS, Orleans CT. Screening for skin cancer. Am J Prev Med 2001;20(Suppl 3):47-58. 41. US Preventive Services Task Force. Screening for skin cancer: recommendations and rationale. Am J Prev Med 2001; 20(Suppl 3):44-6. 42. The Cancer Council Australia. National Cancer Prevention Policy 2001-03. Available at: http://www.cancer.org.au/documents/ National%20Cancer%20Prevention%20Policy.pdf. Accessed July 2, 2003. 43. Hanrahan P, Hersey P, Menzies S, Watson A, D Este C. Examination of the ability of people to identify early changes of melanoma in computer-altered pigmented skin lesions. Arch Dermatol 1997;133:301-11. 44. National Breast Cancer Centre. Position statement. Early detection of breast cancer. Available at: http://www.nbcc.org. au/resources/documents/edp_earlydetectionposition0804.pdf. Accessed November 19, 2004.

792 McPherson et al JAM ACAD DERMATOL MAY 2006 45. Crossing S, Manaszewicz R. Breast self-examination: be alert but not alarmed? Med J Aust 2003;178:646-7. 46. Burton RC, Armstrong BK. Nonmetastasizing melanoma? J Surg Oncol 1998;67:73-6. 47. Burton RC, Armstrong BK. Current melanoma epidemic: a nonmetastasizing form of melanoma? World J Surg 1995;19:330-3. 48. Burton RC. Analysis of public education and the implications with regard to nonprogressive thin melanomas. Curr Opin Oncol 1995;7:170-4. 49. Coory M, Baade PD, Aitken JF, Smithers M, McLeod GRC, Ring I. Trends for in-situ and invasive melanoma in Queensland, Australia, 1982-2002. Cancer Causes Control. In press.