- Email: [email protected]
Melanoma Thickness Trends in the United States, 1988–2006
ORIGINAL ARTICLE
Melanoma Thickness Trends in the United States, 1988–2006 Vincent D. Criscione1,2 and Martin A. Weinstock1,2 Over the past two decades, numerous efforts have been initiated to improve screening and early detection of melanoma both in the United States and worldwide. It is commonly believed that these efforts have contributed to the stabilization of melanoma mortality, and that the proportion of thick melanoma with unfavorable prognosis is on the decline. Data obtained from 17 population-based cancer registries of the Surveillance Epidemiology and End Result (SEER) program of the National Cancer Institute for 1988–2006 were used to examine trends in melanoma tumor thickness. For malignant melanoma cases with recorded thickness, the proportionate distribution among four thickness categories (p1, 1.01–2, 2.01–4, and 44 mm) remained relatively stable over the 19-year study period, however, for melanomas resulting in death, the proportion of thick tumors increased. The most substantial change occurred in the proportion of melanoma in situ, which nearly doubled from 1988 to 2006. Surveillance and early detection efforts in the United States have not resulted in a substantial reduction in the proportion of tumors with prognostically unfavorable thickness. Continued improvement and new methods of screening, especially among demographics with higher incidence of thick tumors, is necessary. Journal of Investigative Dermatology (2010) 130, 793–797; doi:10.1038/jid.2009.328; published online 15 October 2009
INTRODUCTION Malignant melanoma remains one of the most rapidly increasing cancers in the United States. In recent years, the overall mortality has begun to stabilize, and rates have declined among younger age groups (Linos et al., 2009). Better secondary prevention efforts leading to improved detection of early, thinner tumors with lower metastatic potential are thought to be at least partially responsible. Studies conducted in other countries have documented a substantial rise in the proportion of thin tumors over the past two decades (Garbe et al., 2000; Murray et al., 2005; Lipsker et al., 2007; Tejera-Vaquerizo et al., 2008), but few reports have detailed thickness trends in the United States (Dennis, 1999; Demierre et al., 2005; Linos et al., 2009), and only one report directly commented on proportionate changes (Demierre et al., 2005). We are unaware of previous reports that document these trends for fatal melanomas. We used population-based data from the Surveillance Epidemiology and End Results (SEER) program of the National Cancer Institute to examine the trends in tumor thickness for the 1988–2006 period. 1
Departments of Dermatology and Community Health, The Warren Alpert Medical School of Brown University, Providence, Rhode Island, USA and 2 Dermatoepidmiology Unit, VA Medical Center, Providence, Rhode Island, USA Correspondence: Professor Martin A. Weinstock, Dermatoepidemiology Unit111D, VA Medical Center, 830 Chalkstone Avenue, Providence, Rhode Island 02908-4799, USA. E-mail: [email protected] Abbreviations: CI, confidence interval; SEER, Surveillance Epidemiology and End Result Received 17 June 2009; revised 27 August 2009; accepted 28 August 2009; published online 15 October 2009
& 2010 The Society for Investigative Dermatology
RESULTS A total of 153,124 malignant melanoma cases were identified in 17 SEER registries between 1988 and 2006. Of these cases, the majority was histologically classified as melanoma not otherwise specified (47%), superficial spreading melanoma (34%), nodular melanoma (7%) and lentigo maligna melanoma (6%). The thickness was recorded for 85% of cases. The majority of these cases (70%) had tumor thickness p1 mm (Table 1). The proportion of malignant melanoma cases with unknown thickness decreased from 1988–1993 to 2000–2006 (23 vs 13%, 95% confidence interval (CI) of difference 10–12%). Of cases with recorded thickness, the proportion of cases with thickness p1 mm increased from 68 to 71% between 1988–1993 and 2000–2006 (95% CI of difference 2–4%). The proportion of cases with tumor thickness 44 mm also increased from 4.7 to 5.4% (95% CI of difference 0.3–1%). Over the same time periods the proportion of lesions with intermediate thickness (1.01–2 mm and 2.01–4 mm) decreased from 17 to 15% (95% CI of difference 2–3%) and 10 to 9% (95% CI of difference 1–2%), respectively (Figure 1). Similar results were obtained when only cases from the original 9 SEER registries for 1988–2006 were included in the analysis. Among cases of malignant melanoma diagnosed in the 17 SEER registries between 1988 and 2006, 13,993 were determined to be the primary cause of death on death certificate review. Of these fatal melanomas, the majority (54%) carried the histological diagnosis of melanoma not otherwise specified. Nodular melanoma and superficial spreading melanoma made up a similar proportion of fatal cases (18 and 17%, respectively). Only 65% of fatal www.jidonline.org
793
VD Criscione and MA Weinstock Melanoma Thickness Trends
Table 2. Trends in the proportion of fatal melanomas1 classified by tumor thickness, US 1988–2006
Table 1. Proportions of malignant melanoma classified by tumor thickness, US 1988–2006 Tumor thickness
Percentage
No. of cases
Percentage of tumors with known thickness (no. of cases)
p1
70
91,174
Years
1.01–2
16
20,424
1988–1993 30 (649)
27 (579)
27 (579)
16 (352)
36 (1,232)
2.01–4
9
11,702
1994–1999 26 (759)
25 (738)
28 (809)
20 (597)
31 (1,333)
44
5
6,894
Malignant melanomas (mm)
Total
130,194
Fatal melanomas (mm)1 p1
27
2,472
1.01–2
23
2,142
2.01–4
27
2,474
44
22
2,041
Total
9,129
1
Cases of malignant melanoma diagnosed between 1988 and 2006 with melanoma recorded as primary cause of death on death certificate review. Only cases with known thickness at diagnosis were included.
80 70
Percent
60
<=1mm 1.01–2mm 2.01–4mm >4mm
50 40 30 20 10
19 8 19 8 8 19 9 9 19 0 9 19 1 92 19 9 19 3 94 19 9 19 5 9 19 6 97 19 9 19 8 9 20 9 00 20 2001 0 20 2 03 20 0 20 4 05 20 06
0
Year Figure 1. Proportion of malignant melanomas by thickness: US 1988–2006. Note: Only cases with known thickness are included.
melanomas had a recorded tumor thickness. Cases with recorded thickness had a fairly even distribution among the four thickness categories (Table 1). The proportion of fatal melanomas with unknown thickness did not change significantly between 1988–1993 and 2000–2006 (36.33 vs 36.11; P ¼ 0.83, 95% CI of difference 2 to 2%). For fatal melanomas with recorded thickness, the proportion with thickness p1 mm decreased from 30 to 26% between 1988–1993 and 2000–2006 (95% CI of difference 2–6%), whereas the proportion of fatal melanomas with thickness 44 mm increased from 16 to 27% between 1988–1993 and 2000–2006 (95% CI of difference 8–13%) (Table 2). Similar results were obtained when nodular melanoma cases were excluded from the analysis. There were minimal changes in prognosis by thickness group during the study period (Table 3). 794
Journal of Investigative Dermatology (2010), Volume 130
p1 mm 1.01–2 mm 2.01–4 mm 44 mm
2000–2006 26 (1,064) 20 (825)
27 (1,086) 27 (1,092)
Percentage of all tumors Unknown
36 (2,299)
1
Cases of malignant melanoma diagnosed between 1988 and 2006 with melanoma recorded as primary cause of death on death certificate review.
The overall relative survival for melanoma (calculated using study data) stabilizes at approximately 87.8% after 9 years. Hence, almost all deaths resulting from melanoma diagnosed between 1988 and 1997 should have occurred by 2006. Among melanomas diagnosed between 1988 and 1997 with recorded thickness, the proportion of fatal melanomas with thickness p1 mm decreased from 30 to 27% between 1988–1992 and 1993–1997 (P ¼ 0.026; 95% CI of difference 0.4–6.3%). The proportion of melanomas with thickness 44 mm increased from 16 to 19% (P ¼ 0.017; 95% CI of difference 1–6%). Median tumor thickness varied substantially with histological type. Among the most common melanoma subtypes, median thickness was greatest for nodular melanoma (2.3 mm), followed by acral melanoma (1.4 mm), melanoma not otherwise specified (0.6 mm), superficial spreading melanoma (0.6 mm), and lentigo maligna melanoma (0.4 mm). Median tumor thickness for all cases of malignant melanoma decreased from 0.66 to 0.60 mm between 1988–1993 and 2000–2006. Median tumor thickness for fatal melanomas increased from 1.8 mm in 1988–1993 to 2.3 mm in 2000–2006 (Figure 2). A total of 85,852 cases of melanoma in situ were reported to the 17 SEER registries between 1988 and 2006. The proportion of all melanomas that were in situ increased substantially between 1988–1993 and 2000–2006 (25 vs 38%, 95% CI of difference 13–14%) (Figure 3). Several inconsistencies were identified in the SEER data regarding the coding of tumor thickness. Among cases of melanoma in situ, 3,380 (3.4%) were coded as having tumor thickness between 0.01 and 9.9 mm. Of invasive malignant melanoma cases, 1416 (1%) had a recorded tumor thickness of 0 (no mass; no tumor observed). Of these, nearly half (n ¼ 654) were found to be the primary cause of death at follow-up. An inquiry was made to SEER regarding these inconsistencies. DISCUSSION Our analysis of more than 150,000 cases of malignant melanoma demonstrates relatively stable proportions of thick and thin melanoma in the United States over the past two decades. This trend is consistent with our finding of only minimal changes in median tumor thickness and results from earlier population-based studies (Dennis, 1999; Demierre et al., 2005; Linos et al., 2009). Our finding of increased
VD Criscione and MA Weinstock Melanoma Thickness Trends
Table 3. Five-year cause-specific survival1 classified by tumor thickness, US 1988–2006 Five-year survival (95% CI) p1 mm
1.01–2 mm
2.01–4 mm
44 mm
1988–1992
96.8 (96.4–97.2)
86.0 (84.3–87.5)
72.3 (69.4–75.0)
55.4 (50.7–59.8)
1993–1997
97.2 (96.9–97.5)
87.8 (86.6–89.0)
71.3 (69.0–73.4)
56.8 (53.3–60.1)
1998–2002
97.1 (96.9–97.4)
89.1 (88.2–89.9)
74.4 (72.8–76.0)
58.1 (55.5–60.5)
Year of diagnosis
CI, confidence interval. 1 Cause-specific survival is the probability of not dying from melanoma. It was calculated using the Kaplan–Meier method with cause of death defined by the primary site, and the cause of death recorded on the death certificate.
Median tumor thickness (mm)
6 All malignant melanomas
5
Fatal melanomas
4 3 2 1
19 8 19 8 89 19 9 19 0 91 19 9 19 2 9 19 3 94 19 9 19 5 9 19 6 97 19 9 19 8 9 20 9 0 20 0 0 20 1 02 20 0 20 3 04 20 0 20 5 06
0 Year Figure 2. Median tumor thickness: US 1988–2006.
45 40 35 Percent
30 25 20 15 10 5 19 8 19 8 89 19 9 19 0 9 19 1 9 19 2 9 19 3 94 19 9 19 5 9 19 6 9 19 7 98 19 9 20 9 0 20 0 0 20 1 0 20 2 0 20 3 0 20 4 0 20 5 06
0
Year Figure 3. Proportion of all melanomas that were melanoma in situ: US 1988–2006.
proportions of thick tumors among fatal melanomas after 2003 is probably due to deaths from thinner melanoma diagnosed during the later years of the study occurring after the study period, hence not reported herein. Thicker tumors tend to result in more rapid demise and are likely to be disproportionately represented among deaths occurring shortly after diagnosis. These trends are striking given elevated public awareness and improved screening efforts aimed at detecting early,
thinner lesions over the past two decades. Beginning in 1985, skin self examinations were increasingly promoted after the development of the ABCD (asymmetry, border, color, diameter) acronym (Heymann, 2007). At the same time, dermatologist-led national screening and educational efforts sponsored by the American Academy of Dermatology were initiated, and have provided important educational messages about sun protection and early detection (Geller et al., 2003). Improved understanding of host risk factors has helped to target screening efforts on high-risk groups, and the utilization of new technologies such as dermoscopy has improved the diagnostic accuracy (Geller et al., 2007). Internationally, such interventions have had a larger impact in some countries than in others. Population-based studies from Scotland (Mackie et al., 2002; Murray et al., 2005), France (Lipsker et al., 2007), Southern Germany (Garbe et al., 2000), and Spain (Tejera-Vaquerizo et al., 2008) demonstrate a substantial rise in the proportion of thin tumors over the past two decades, whereas data from Australia show proportions and trends comparable to those seen in the United States (Garbe et al., 2000). For example, Garbe et al. (2000) observed that the proportion of thin (p0.75 mm) tumors among cases of melanoma recorded by the Queensland Melanoma register increased from 61.9 to 63.6% between 1986 and 1996. During this same period, the proportion of thin (p0.75 mm) melanomas among cases recorded by the Central Malignant Melanoma Register of the German Dermatological Society increased from 29.8 to 46.4%. Comparably, the proportion of thin (o1.5 mm) tumors recorded by the Scottish Melanoma Group increased from 38 to 60% between 1985 and 2004 (Murray et al., 2005). It seems that populations experiencing the greatest rise in the proportion of thin tumors are those that began with substantially lower proportions at baseline. Current screening interventions and elevated public awareness are effective in these areas but produce limited returns in areas with longstanding screening and awareness campaigns. Although current screening efforts have not had a substantial effect on the early detection of thinner malignant lesions in the United States, they have certainly had a significant impact on the detection of non-invasive melanoma, the proportion of which doubled over the study period. A similar rise in the incidence of melanoma in situ has been reported in Australia (Coory et al., 2006), and in California among an employed population undergoing increased www.jidonline.org
795
VD Criscione and MA Weinstock Melanoma Thickness Trends
screening (Schneider et al., 2008). There is considerable uncertainty about the invasive potential of these lesions, which reflects the dearth of study of their transformation. In a survey of 1200 dermatologists, 83.6% considered melanoma in situ to be a true malignancy, but 27.8% were unable to estimate the proportion of untreated lesions that would progress to invasive growth and 43.9% were unsure of the metastatic potential (Charles et al., 2005). Unlike the trends seen with cervical cancer screening, the incidence of invasive melanoma continues to rise and mortality remains stable despite large increases in the identification of noninvasive lesions. This trend supports the argument that increased screening efforts are leading to greater identification of biologically inert lesions that have always been present, but were not previously diagnosed as melanoma (Burton and Armstrong, 1998). Our results demonstrate a clear need for new screening strategies. In addition, the expansion of current screening techniques, such as the full body skin examination and thorough skin self-examination are necessary. In a national health survey of over 30,000 patients, the reported rate of full body skin examination was only 21% (Santmyire et al., 2001). Other studies found that the rate of thorough skin selfexamination ranged from 9 to 18% only (Weinstock et al., 1999, 2004). Several shortcomings of the US healthcare system provide barriers to expanded screening. Inequalities in access to healthcare limit physician screening among patients of low socioeconomic status and may result in increased incidence of thick, late stage tumors and decreased survival (Eide et al., 2009; Linos et al., 2009). Among the insured patients, physician factors that inhibit screening include shortages of primary care physicians, limited time for patient encounters, lack of incentives for preventative care, and absence of skin examination training (Moore et al., 2006). Even with additional time and better training, preventative strategies will remain underutilized without clear recommendations from authoritative bodies backed by evidence showing the benefits of early detection of skin cancer (Lee and Weinstock, 2009). Although anticipated comprehensive healthcare reform may help to address some of these issues, rising incidence, non-declining overall mortality and relatively unchanging thickness distributions demand immediate action. Geller et al. (2006) recently outlined a proposal for a targeted national melanoma-screening program consisting of policy changes, expanded public outreach, education of clinicians and legislative advocacy. To justify such an undertaking, studies demonstrating the benefits of population-wide screenings are necessary and may be forthcoming (Koh, 2007; Geller, 2009). One screening and educational program initiated among employees of the Lawrence Livermore National Laboratory in California resulted in a 69% reduction in crude incidence of thick (40.75 mm) tumors (Schneider et al., 2008). Additional data from a large randomized screening trial in Queensland, Australia, demonstrated high specificity for full body skin examination (Aitken et al., 2006a), and a significant rise in the number of full body skin examinations among intervention groups (Aitken et al., 796
Journal of Investigative Dermatology (2010), Volume 130
2006b). Recently, investigators formed the Melanoma Screening Group and have begun working towards the creation of a similar melanoma early detection trial in the United States (Geller, 2009). In preparation for new screening modalities, it is important to have a data collection system that can accurately measure population-wide effects. Our study identifies several problems within SEER data that limit this ability. First, the underreporting of tumor thickness, although improving, remains high, especially among fatal cases. Second, ambiguity in thickness codes and inconsistencies in the reporting of tumor thickness are present in SEER. We identified over 1,000 cases of malignant, invasive melanoma that were recorded as having a thickness of 0, defined as ‘‘no mass; no tumor found.’’ Similarly, several thousand cases of in situ melanoma were recorded as having a non-zero thickness. The upper limit of recording thickness in SEER is 9.90 mm, with no option for recording thickness of tumors greater than this cutoff. These problems limit the analysis of melanoma thickness trends and should be rectified. Additional limitations of SEER data have been documented and include absence of independent verification of diagnoses (Merlino et al., 1997), lack of case detail, underregistration (Cockburn et al., 2008) and delayed reporting (Clegg et al., 2002). Our population-based study, the largest assessment to date, provides comprehensive analysis of tumor thickness trends in the United States and describes patterns among fatal melanomas that have not been previously reported. These data highlight the need for new and expanded melanoma screening modalities and improvements in the collection and recording of data used to measure their effectiveness. MATERIALS AND METHODS Data were obtained from 17 population-based cancer registries of the SEER program of the US National Cancer Institute (SEER, 2008). The study period (1988–2006) includes all years in which tumor thickness was consistently recorded. Data from the original 9 SEER registries were available for the entire study period. Four additional registries (Los Angelos and San Jose, California; Alaska; and rural Georgia) added data for the last 15 years of the study period, and another four registries (Greater California excluding San Francisco, Los Angeles and San Jose; Kentucky; Louisiana; and New Jersey) provided data for the last 6 years of the period. Together, these seventeen SEER registries represent approximately 26% of the US population. For study purposes, cases of malignant melanoma were defined as those with International Classification of Diseases for Oncology (ICD-O) histological types melanoma not otherwise specified (code 8720), nodular melanoma (code 8721), lentigo maligna melanoma (code 8742), superficial spreading melanoma (code 8743), acral lentiginous melanoma (code 8744), other melanoma subcategories (codes 8777, 8723, 8730, 8740, 8741, 8745, 8746, 8761, 8770, 8771, 8772), ICD-O behavior type malignant (code 3), and primary site of skin (codes 440–449). Tumor thickness was stratified according to the American Joint Committee on Cancer thickness categories (p1, 1.01–2, 2.01–4, 44 mm). Cases of malignant melanoma with ‘‘no mass or tumor found’’ (extent of disease code 0) were not included in the analysis (see last paragraph of Results section for further information on inconsistencies
VD Criscione and MA Weinstock Melanoma Thickness Trends
observed). Fatal melanomas were defined as cases of malignant melanoma diagnosed between 1988 and 2006, and who died by 2006, with melanoma recorded as the primary cause of death on death certificate review. These deaths were identified in SEER by cause of death to site recode (code 25010). Survival was calculated with SEER*Stat using the Kaplan–Meier method. Cause-specific survival is the probability of not dying from melanoma. Melanoma in situ were identified by ICD-O behavior code in situ (code 2). Only cases with unrecorded thickness (extent of disease code 999) or no mass found (extent of disease code 0) were included in the analysis of in situ cases (see last paragraph of Results section for additional details). Data were analyzed using Stata SE version 8 (StataCorp, College Station, TX). CI values were calculated using proportion tests in Stata SE. Given the large sample size, all differences in proportions were statistically significant (Po0.001) unless otherwise specified.
Garbe C, McLeod G, Buettner P (2000) Time trends of cutaneous melanoma in Queensland, Australia and Central Europe. Cancer 89:1269–78
CONFLICT OF INTEREST
Lee K, Weinstock M (2009) Melanoma is up: are we up to this challenge? J Invest Dermatol 129:1604–6
The authors state no conflict of interest.
ACKNOWLEDGMENTS The study was supported by the Department of Veterans Affairs Office of Research and Development Co-operative Studies Program and by grants R01CA106592, R01CA106807, and R25CA087972 from the National Institutes of Health (MAW).
REFERENCES Aitken J, Janda M, Elwood M, Youl P, Ring I, Lowe J (2006a) Clinical outcomes from skin screening clinics within a community-based melanoma screening program. J Am Acad Dermatol 54:105–14 Aitken J, Youl PMJ, Lowe J, Ring I, Elwood M (2006b) Increase in skin cancer screening during a community-based randomized intervention trial. Int J Cancer 118:1010–6 Burton R, Armstrong B (1998) Non-metastasizing melanoma? J Surg Oncol 67:73–6 Charles C, Yee V, Dusza S, Marghoob A, Oliveria S, Kopf A et al. (2005) Variation in the diagnosis, treatment, and management of melanoma in situ: a survey of US dermatologists. Arch Dermatol 141:723–9 Clegg L, Feuer E, Midthune D, Fay M, Hankey B (2002) Impact of reporting delay and reporting error on cancer incidence rates and trends. J Natl Cancer Inst 94:1537–45 Cockburn M, Swetter SM, Peng D, Keegan T, Deapen D, Clarke CA (2008) Melanoma underreporting: why does it happen, how big is the problem. J Am Acad Dermatol 59:1081–5 Coory M, Baade P, Aitken J, Smithers M, Roderick G, Mcleod C et al. (2006) Trends for in situ and invasive melanoma in Queensland, Australia, 1982–2002. Cancer Causes Control 17:21–7 Demierre MF, Chung C, Miller D, Geller A (2005) Early detection of thick melanomas in the United States: beware of the nodular subtype. Arch Dermatol 141:745–50 Dennis LK (1999) Analysis of the melanoma epidemic, both apparent and real: data from the 1973 through 1994 surveillance, epidemiology, and end results program registry. Arch Dermatol 135:275–80 Eide M, Weinstock M, Clark M (2009) Demographic and socioeconomic predictors of melanoma prognosis in the United States. J Health Care Poor Underserved 20:227–45
Geller A (2009) Educational and screening campaigns to reduce deaths from melanoma. Hematol Oncol Clin North Am 23:515–27 Geller A, Miller D, Swetter SM, Demierre MF, Gilchrest B (2006) A call for the development and implementation of a targeted national melanoma screening program. Arch Dermatol 142:504–7 Geller A, Zhang Z, Sober AJ, Halpern A, Weinstock M, Daniels S et al. (2003) The first 15 years of the American Academy of Dermatology skin cancer screening programs: 1985–1999. J Am Acad Dermatol 48:34–41 Geller J, Swetter SM, Leyson J, Miller D, Brooks K, Geller A (2007) Screening, early detection, and trends for melanoma: current status (2000–2006) and future directions. J Am Acad Dermatol 57:555–72 Heymann WR (2007) Screening for melanoma. J Am Acad Dermatol 56:144–5 Koh H (2007) Melanoma screening: focusing the public health journey. Arch Dermatol 143:101–3
Linos E, Swetter SM, Cockburn MG, Colditz MG, Clarke CA (2009) Increasing burden of melanoma in the United States. J Invest Dermatol 129:1666–74 Lipsker D, Engel F, Cribier B, Velten M, Hedelin G (2007) Trends in melanoma epidemiology suggest three different types of melanoma. Br J Dermatol 157:338–43 Mackie R, Bray C, Hole D, Morris A, Nicolson M, Evans A et al. (2002) Incidence of and survival from malignant melanoma in Scotland: an epidemiologic study. Lancet 360:587–91 Merlino L, Sullivan K, Whitaker D, Lynch C (1997) The independent pathology laboratory as a reporting source for cutaneous melanoma incidence in Iowa, 1977–1994. J Am Acad Dermatol 37:578–85 Moore M, Geller A, Zhang Z, Hayes B, Bergstrom K, Graves J et al. (2006) Skin cancer examination teaching in US medical education. Arch Dermatol 142:439–44 Murray CS, Stockton DL, Doherty VR (2005) Thick melanoma: the challenge persists. Br J Dermatol 152:104–9 Santmyire B, Feldman S, Fleischer A (2001) Lifestyle high-risk behaviors and demographics may predict the level of participation in sun-protection behaviors and skin cancer primary prevention in the United States: results of the 1998 National Health Interview Survey. Cancer 92:1315–24 Schneider J, Moore D, Mendelsohn M (2008) Screening program reduced melanoma mortality at the Lawrence Livermore National Laboratory, 1984 to 1996. J Am Acad Dermatol 58:741–9 SEER (2008) Surveillance, Epidemiology, and End Results Program. Available at www.seer.cancer.gov. Limited-use data (1973–2006), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2009, based on the November 2008 submission. Tejera-Vaquerizo A, Mendiola-Fernandez M, Fernandez-Orland A, HerreraCeballos E (2008) Thick melanoma: the problem continues. J Eur Acad Dermatol Venereol 22:575–9 Weinstock M, Martin R, Risica P, Bwerwick M, Lasater T, Rakowski W et al. (1999) Thorough skin examination for the early detection of melanoma. Am J Prev Med 17:169–75 Weinstock M, Risica P, Martin R, Rakowski W, Smith K, Berwick M et al. (2004) Reliability of assessment and circumstances of performance of thorough skin self-examination for the early detection of melanoma in the Check-It-Out Project. Prev Med 38:761–5
www.jidonline.org
797
Melanoma Thickness Trends in the United States, 1988–2006 Vincent D. Criscione1,2 and Martin A. Weinstock1,2 Over the past two decades, numerous efforts have been initiated to improve screening and early detection of melanoma both in the United States and worldwide. It is commonly believed that these efforts have contributed to the stabilization of melanoma mortality, and that the proportion of thick melanoma with unfavorable prognosis is on the decline. Data obtained from 17 population-based cancer registries of the Surveillance Epidemiology and End Result (SEER) program of the National Cancer Institute for 1988–2006 were used to examine trends in melanoma tumor thickness. For malignant melanoma cases with recorded thickness, the proportionate distribution among four thickness categories (p1, 1.01–2, 2.01–4, and 44 mm) remained relatively stable over the 19-year study period, however, for melanomas resulting in death, the proportion of thick tumors increased. The most substantial change occurred in the proportion of melanoma in situ, which nearly doubled from 1988 to 2006. Surveillance and early detection efforts in the United States have not resulted in a substantial reduction in the proportion of tumors with prognostically unfavorable thickness. Continued improvement and new methods of screening, especially among demographics with higher incidence of thick tumors, is necessary. Journal of Investigative Dermatology (2010) 130, 793–797; doi:10.1038/jid.2009.328; published online 15 October 2009
INTRODUCTION Malignant melanoma remains one of the most rapidly increasing cancers in the United States. In recent years, the overall mortality has begun to stabilize, and rates have declined among younger age groups (Linos et al., 2009). Better secondary prevention efforts leading to improved detection of early, thinner tumors with lower metastatic potential are thought to be at least partially responsible. Studies conducted in other countries have documented a substantial rise in the proportion of thin tumors over the past two decades (Garbe et al., 2000; Murray et al., 2005; Lipsker et al., 2007; Tejera-Vaquerizo et al., 2008), but few reports have detailed thickness trends in the United States (Dennis, 1999; Demierre et al., 2005; Linos et al., 2009), and only one report directly commented on proportionate changes (Demierre et al., 2005). We are unaware of previous reports that document these trends for fatal melanomas. We used population-based data from the Surveillance Epidemiology and End Results (SEER) program of the National Cancer Institute to examine the trends in tumor thickness for the 1988–2006 period. 1
Departments of Dermatology and Community Health, The Warren Alpert Medical School of Brown University, Providence, Rhode Island, USA and 2 Dermatoepidmiology Unit, VA Medical Center, Providence, Rhode Island, USA Correspondence: Professor Martin A. Weinstock, Dermatoepidemiology Unit111D, VA Medical Center, 830 Chalkstone Avenue, Providence, Rhode Island 02908-4799, USA. E-mail: [email protected] Abbreviations: CI, confidence interval; SEER, Surveillance Epidemiology and End Result Received 17 June 2009; revised 27 August 2009; accepted 28 August 2009; published online 15 October 2009
& 2010 The Society for Investigative Dermatology
RESULTS A total of 153,124 malignant melanoma cases were identified in 17 SEER registries between 1988 and 2006. Of these cases, the majority was histologically classified as melanoma not otherwise specified (47%), superficial spreading melanoma (34%), nodular melanoma (7%) and lentigo maligna melanoma (6%). The thickness was recorded for 85% of cases. The majority of these cases (70%) had tumor thickness p1 mm (Table 1). The proportion of malignant melanoma cases with unknown thickness decreased from 1988–1993 to 2000–2006 (23 vs 13%, 95% confidence interval (CI) of difference 10–12%). Of cases with recorded thickness, the proportion of cases with thickness p1 mm increased from 68 to 71% between 1988–1993 and 2000–2006 (95% CI of difference 2–4%). The proportion of cases with tumor thickness 44 mm also increased from 4.7 to 5.4% (95% CI of difference 0.3–1%). Over the same time periods the proportion of lesions with intermediate thickness (1.01–2 mm and 2.01–4 mm) decreased from 17 to 15% (95% CI of difference 2–3%) and 10 to 9% (95% CI of difference 1–2%), respectively (Figure 1). Similar results were obtained when only cases from the original 9 SEER registries for 1988–2006 were included in the analysis. Among cases of malignant melanoma diagnosed in the 17 SEER registries between 1988 and 2006, 13,993 were determined to be the primary cause of death on death certificate review. Of these fatal melanomas, the majority (54%) carried the histological diagnosis of melanoma not otherwise specified. Nodular melanoma and superficial spreading melanoma made up a similar proportion of fatal cases (18 and 17%, respectively). Only 65% of fatal www.jidonline.org
793
VD Criscione and MA Weinstock Melanoma Thickness Trends
Table 2. Trends in the proportion of fatal melanomas1 classified by tumor thickness, US 1988–2006
Table 1. Proportions of malignant melanoma classified by tumor thickness, US 1988–2006 Tumor thickness
Percentage
No. of cases
Percentage of tumors with known thickness (no. of cases)
p1
70
91,174
Years
1.01–2
16
20,424
1988–1993 30 (649)
27 (579)
27 (579)
16 (352)
36 (1,232)
2.01–4
9
11,702
1994–1999 26 (759)
25 (738)
28 (809)
20 (597)
31 (1,333)
44
5
6,894
Malignant melanomas (mm)
Total
130,194
Fatal melanomas (mm)1 p1
27
2,472
1.01–2
23
2,142
2.01–4
27
2,474
44
22
2,041
Total
9,129
1
Cases of malignant melanoma diagnosed between 1988 and 2006 with melanoma recorded as primary cause of death on death certificate review. Only cases with known thickness at diagnosis were included.
80 70
Percent
60
<=1mm 1.01–2mm 2.01–4mm >4mm
50 40 30 20 10
19 8 19 8 8 19 9 9 19 0 9 19 1 92 19 9 19 3 94 19 9 19 5 9 19 6 97 19 9 19 8 9 20 9 00 20 2001 0 20 2 03 20 0 20 4 05 20 06
0
Year Figure 1. Proportion of malignant melanomas by thickness: US 1988–2006. Note: Only cases with known thickness are included.
melanomas had a recorded tumor thickness. Cases with recorded thickness had a fairly even distribution among the four thickness categories (Table 1). The proportion of fatal melanomas with unknown thickness did not change significantly between 1988–1993 and 2000–2006 (36.33 vs 36.11; P ¼ 0.83, 95% CI of difference 2 to 2%). For fatal melanomas with recorded thickness, the proportion with thickness p1 mm decreased from 30 to 26% between 1988–1993 and 2000–2006 (95% CI of difference 2–6%), whereas the proportion of fatal melanomas with thickness 44 mm increased from 16 to 27% between 1988–1993 and 2000–2006 (95% CI of difference 8–13%) (Table 2). Similar results were obtained when nodular melanoma cases were excluded from the analysis. There were minimal changes in prognosis by thickness group during the study period (Table 3). 794
Journal of Investigative Dermatology (2010), Volume 130
p1 mm 1.01–2 mm 2.01–4 mm 44 mm
2000–2006 26 (1,064) 20 (825)
27 (1,086) 27 (1,092)
Percentage of all tumors Unknown
36 (2,299)
1
Cases of malignant melanoma diagnosed between 1988 and 2006 with melanoma recorded as primary cause of death on death certificate review.
The overall relative survival for melanoma (calculated using study data) stabilizes at approximately 87.8% after 9 years. Hence, almost all deaths resulting from melanoma diagnosed between 1988 and 1997 should have occurred by 2006. Among melanomas diagnosed between 1988 and 1997 with recorded thickness, the proportion of fatal melanomas with thickness p1 mm decreased from 30 to 27% between 1988–1992 and 1993–1997 (P ¼ 0.026; 95% CI of difference 0.4–6.3%). The proportion of melanomas with thickness 44 mm increased from 16 to 19% (P ¼ 0.017; 95% CI of difference 1–6%). Median tumor thickness varied substantially with histological type. Among the most common melanoma subtypes, median thickness was greatest for nodular melanoma (2.3 mm), followed by acral melanoma (1.4 mm), melanoma not otherwise specified (0.6 mm), superficial spreading melanoma (0.6 mm), and lentigo maligna melanoma (0.4 mm). Median tumor thickness for all cases of malignant melanoma decreased from 0.66 to 0.60 mm between 1988–1993 and 2000–2006. Median tumor thickness for fatal melanomas increased from 1.8 mm in 1988–1993 to 2.3 mm in 2000–2006 (Figure 2). A total of 85,852 cases of melanoma in situ were reported to the 17 SEER registries between 1988 and 2006. The proportion of all melanomas that were in situ increased substantially between 1988–1993 and 2000–2006 (25 vs 38%, 95% CI of difference 13–14%) (Figure 3). Several inconsistencies were identified in the SEER data regarding the coding of tumor thickness. Among cases of melanoma in situ, 3,380 (3.4%) were coded as having tumor thickness between 0.01 and 9.9 mm. Of invasive malignant melanoma cases, 1416 (1%) had a recorded tumor thickness of 0 (no mass; no tumor observed). Of these, nearly half (n ¼ 654) were found to be the primary cause of death at follow-up. An inquiry was made to SEER regarding these inconsistencies. DISCUSSION Our analysis of more than 150,000 cases of malignant melanoma demonstrates relatively stable proportions of thick and thin melanoma in the United States over the past two decades. This trend is consistent with our finding of only minimal changes in median tumor thickness and results from earlier population-based studies (Dennis, 1999; Demierre et al., 2005; Linos et al., 2009). Our finding of increased
VD Criscione and MA Weinstock Melanoma Thickness Trends
Table 3. Five-year cause-specific survival1 classified by tumor thickness, US 1988–2006 Five-year survival (95% CI) p1 mm
1.01–2 mm
2.01–4 mm
44 mm
1988–1992
96.8 (96.4–97.2)
86.0 (84.3–87.5)
72.3 (69.4–75.0)
55.4 (50.7–59.8)
1993–1997
97.2 (96.9–97.5)
87.8 (86.6–89.0)
71.3 (69.0–73.4)
56.8 (53.3–60.1)
1998–2002
97.1 (96.9–97.4)
89.1 (88.2–89.9)
74.4 (72.8–76.0)
58.1 (55.5–60.5)
Year of diagnosis
CI, confidence interval. 1 Cause-specific survival is the probability of not dying from melanoma. It was calculated using the Kaplan–Meier method with cause of death defined by the primary site, and the cause of death recorded on the death certificate.
Median tumor thickness (mm)
6 All malignant melanomas
5
Fatal melanomas
4 3 2 1
19 8 19 8 89 19 9 19 0 91 19 9 19 2 9 19 3 94 19 9 19 5 9 19 6 97 19 9 19 8 9 20 9 0 20 0 0 20 1 02 20 0 20 3 04 20 0 20 5 06
0 Year Figure 2. Median tumor thickness: US 1988–2006.
45 40 35 Percent
30 25 20 15 10 5 19 8 19 8 89 19 9 19 0 9 19 1 9 19 2 9 19 3 94 19 9 19 5 9 19 6 9 19 7 98 19 9 20 9 0 20 0 0 20 1 0 20 2 0 20 3 0 20 4 0 20 5 06
0
Year Figure 3. Proportion of all melanomas that were melanoma in situ: US 1988–2006.
proportions of thick tumors among fatal melanomas after 2003 is probably due to deaths from thinner melanoma diagnosed during the later years of the study occurring after the study period, hence not reported herein. Thicker tumors tend to result in more rapid demise and are likely to be disproportionately represented among deaths occurring shortly after diagnosis. These trends are striking given elevated public awareness and improved screening efforts aimed at detecting early,
thinner lesions over the past two decades. Beginning in 1985, skin self examinations were increasingly promoted after the development of the ABCD (asymmetry, border, color, diameter) acronym (Heymann, 2007). At the same time, dermatologist-led national screening and educational efforts sponsored by the American Academy of Dermatology were initiated, and have provided important educational messages about sun protection and early detection (Geller et al., 2003). Improved understanding of host risk factors has helped to target screening efforts on high-risk groups, and the utilization of new technologies such as dermoscopy has improved the diagnostic accuracy (Geller et al., 2007). Internationally, such interventions have had a larger impact in some countries than in others. Population-based studies from Scotland (Mackie et al., 2002; Murray et al., 2005), France (Lipsker et al., 2007), Southern Germany (Garbe et al., 2000), and Spain (Tejera-Vaquerizo et al., 2008) demonstrate a substantial rise in the proportion of thin tumors over the past two decades, whereas data from Australia show proportions and trends comparable to those seen in the United States (Garbe et al., 2000). For example, Garbe et al. (2000) observed that the proportion of thin (p0.75 mm) tumors among cases of melanoma recorded by the Queensland Melanoma register increased from 61.9 to 63.6% between 1986 and 1996. During this same period, the proportion of thin (p0.75 mm) melanomas among cases recorded by the Central Malignant Melanoma Register of the German Dermatological Society increased from 29.8 to 46.4%. Comparably, the proportion of thin (o1.5 mm) tumors recorded by the Scottish Melanoma Group increased from 38 to 60% between 1985 and 2004 (Murray et al., 2005). It seems that populations experiencing the greatest rise in the proportion of thin tumors are those that began with substantially lower proportions at baseline. Current screening interventions and elevated public awareness are effective in these areas but produce limited returns in areas with longstanding screening and awareness campaigns. Although current screening efforts have not had a substantial effect on the early detection of thinner malignant lesions in the United States, they have certainly had a significant impact on the detection of non-invasive melanoma, the proportion of which doubled over the study period. A similar rise in the incidence of melanoma in situ has been reported in Australia (Coory et al., 2006), and in California among an employed population undergoing increased www.jidonline.org
795
VD Criscione and MA Weinstock Melanoma Thickness Trends
screening (Schneider et al., 2008). There is considerable uncertainty about the invasive potential of these lesions, which reflects the dearth of study of their transformation. In a survey of 1200 dermatologists, 83.6% considered melanoma in situ to be a true malignancy, but 27.8% were unable to estimate the proportion of untreated lesions that would progress to invasive growth and 43.9% were unsure of the metastatic potential (Charles et al., 2005). Unlike the trends seen with cervical cancer screening, the incidence of invasive melanoma continues to rise and mortality remains stable despite large increases in the identification of noninvasive lesions. This trend supports the argument that increased screening efforts are leading to greater identification of biologically inert lesions that have always been present, but were not previously diagnosed as melanoma (Burton and Armstrong, 1998). Our results demonstrate a clear need for new screening strategies. In addition, the expansion of current screening techniques, such as the full body skin examination and thorough skin self-examination are necessary. In a national health survey of over 30,000 patients, the reported rate of full body skin examination was only 21% (Santmyire et al., 2001). Other studies found that the rate of thorough skin selfexamination ranged from 9 to 18% only (Weinstock et al., 1999, 2004). Several shortcomings of the US healthcare system provide barriers to expanded screening. Inequalities in access to healthcare limit physician screening among patients of low socioeconomic status and may result in increased incidence of thick, late stage tumors and decreased survival (Eide et al., 2009; Linos et al., 2009). Among the insured patients, physician factors that inhibit screening include shortages of primary care physicians, limited time for patient encounters, lack of incentives for preventative care, and absence of skin examination training (Moore et al., 2006). Even with additional time and better training, preventative strategies will remain underutilized without clear recommendations from authoritative bodies backed by evidence showing the benefits of early detection of skin cancer (Lee and Weinstock, 2009). Although anticipated comprehensive healthcare reform may help to address some of these issues, rising incidence, non-declining overall mortality and relatively unchanging thickness distributions demand immediate action. Geller et al. (2006) recently outlined a proposal for a targeted national melanoma-screening program consisting of policy changes, expanded public outreach, education of clinicians and legislative advocacy. To justify such an undertaking, studies demonstrating the benefits of population-wide screenings are necessary and may be forthcoming (Koh, 2007; Geller, 2009). One screening and educational program initiated among employees of the Lawrence Livermore National Laboratory in California resulted in a 69% reduction in crude incidence of thick (40.75 mm) tumors (Schneider et al., 2008). Additional data from a large randomized screening trial in Queensland, Australia, demonstrated high specificity for full body skin examination (Aitken et al., 2006a), and a significant rise in the number of full body skin examinations among intervention groups (Aitken et al., 796
Journal of Investigative Dermatology (2010), Volume 130
2006b). Recently, investigators formed the Melanoma Screening Group and have begun working towards the creation of a similar melanoma early detection trial in the United States (Geller, 2009). In preparation for new screening modalities, it is important to have a data collection system that can accurately measure population-wide effects. Our study identifies several problems within SEER data that limit this ability. First, the underreporting of tumor thickness, although improving, remains high, especially among fatal cases. Second, ambiguity in thickness codes and inconsistencies in the reporting of tumor thickness are present in SEER. We identified over 1,000 cases of malignant, invasive melanoma that were recorded as having a thickness of 0, defined as ‘‘no mass; no tumor found.’’ Similarly, several thousand cases of in situ melanoma were recorded as having a non-zero thickness. The upper limit of recording thickness in SEER is 9.90 mm, with no option for recording thickness of tumors greater than this cutoff. These problems limit the analysis of melanoma thickness trends and should be rectified. Additional limitations of SEER data have been documented and include absence of independent verification of diagnoses (Merlino et al., 1997), lack of case detail, underregistration (Cockburn et al., 2008) and delayed reporting (Clegg et al., 2002). Our population-based study, the largest assessment to date, provides comprehensive analysis of tumor thickness trends in the United States and describes patterns among fatal melanomas that have not been previously reported. These data highlight the need for new and expanded melanoma screening modalities and improvements in the collection and recording of data used to measure their effectiveness. MATERIALS AND METHODS Data were obtained from 17 population-based cancer registries of the SEER program of the US National Cancer Institute (SEER, 2008). The study period (1988–2006) includes all years in which tumor thickness was consistently recorded. Data from the original 9 SEER registries were available for the entire study period. Four additional registries (Los Angelos and San Jose, California; Alaska; and rural Georgia) added data for the last 15 years of the study period, and another four registries (Greater California excluding San Francisco, Los Angeles and San Jose; Kentucky; Louisiana; and New Jersey) provided data for the last 6 years of the period. Together, these seventeen SEER registries represent approximately 26% of the US population. For study purposes, cases of malignant melanoma were defined as those with International Classification of Diseases for Oncology (ICD-O) histological types melanoma not otherwise specified (code 8720), nodular melanoma (code 8721), lentigo maligna melanoma (code 8742), superficial spreading melanoma (code 8743), acral lentiginous melanoma (code 8744), other melanoma subcategories (codes 8777, 8723, 8730, 8740, 8741, 8745, 8746, 8761, 8770, 8771, 8772), ICD-O behavior type malignant (code 3), and primary site of skin (codes 440–449). Tumor thickness was stratified according to the American Joint Committee on Cancer thickness categories (p1, 1.01–2, 2.01–4, 44 mm). Cases of malignant melanoma with ‘‘no mass or tumor found’’ (extent of disease code 0) were not included in the analysis (see last paragraph of Results section for further information on inconsistencies
VD Criscione and MA Weinstock Melanoma Thickness Trends
observed). Fatal melanomas were defined as cases of malignant melanoma diagnosed between 1988 and 2006, and who died by 2006, with melanoma recorded as the primary cause of death on death certificate review. These deaths were identified in SEER by cause of death to site recode (code 25010). Survival was calculated with SEER*Stat using the Kaplan–Meier method. Cause-specific survival is the probability of not dying from melanoma. Melanoma in situ were identified by ICD-O behavior code in situ (code 2). Only cases with unrecorded thickness (extent of disease code 999) or no mass found (extent of disease code 0) were included in the analysis of in situ cases (see last paragraph of Results section for additional details). Data were analyzed using Stata SE version 8 (StataCorp, College Station, TX). CI values were calculated using proportion tests in Stata SE. Given the large sample size, all differences in proportions were statistically significant (Po0.001) unless otherwise specified.
Garbe C, McLeod G, Buettner P (2000) Time trends of cutaneous melanoma in Queensland, Australia and Central Europe. Cancer 89:1269–78
CONFLICT OF INTEREST
Lee K, Weinstock M (2009) Melanoma is up: are we up to this challenge? J Invest Dermatol 129:1604–6
The authors state no conflict of interest.
ACKNOWLEDGMENTS The study was supported by the Department of Veterans Affairs Office of Research and Development Co-operative Studies Program and by grants R01CA106592, R01CA106807, and R25CA087972 from the National Institutes of Health (MAW).
REFERENCES Aitken J, Janda M, Elwood M, Youl P, Ring I, Lowe J (2006a) Clinical outcomes from skin screening clinics within a community-based melanoma screening program. J Am Acad Dermatol 54:105–14 Aitken J, Youl PMJ, Lowe J, Ring I, Elwood M (2006b) Increase in skin cancer screening during a community-based randomized intervention trial. Int J Cancer 118:1010–6 Burton R, Armstrong B (1998) Non-metastasizing melanoma? J Surg Oncol 67:73–6 Charles C, Yee V, Dusza S, Marghoob A, Oliveria S, Kopf A et al. (2005) Variation in the diagnosis, treatment, and management of melanoma in situ: a survey of US dermatologists. Arch Dermatol 141:723–9 Clegg L, Feuer E, Midthune D, Fay M, Hankey B (2002) Impact of reporting delay and reporting error on cancer incidence rates and trends. J Natl Cancer Inst 94:1537–45 Cockburn M, Swetter SM, Peng D, Keegan T, Deapen D, Clarke CA (2008) Melanoma underreporting: why does it happen, how big is the problem. J Am Acad Dermatol 59:1081–5 Coory M, Baade P, Aitken J, Smithers M, Roderick G, Mcleod C et al. (2006) Trends for in situ and invasive melanoma in Queensland, Australia, 1982–2002. Cancer Causes Control 17:21–7 Demierre MF, Chung C, Miller D, Geller A (2005) Early detection of thick melanomas in the United States: beware of the nodular subtype. Arch Dermatol 141:745–50 Dennis LK (1999) Analysis of the melanoma epidemic, both apparent and real: data from the 1973 through 1994 surveillance, epidemiology, and end results program registry. Arch Dermatol 135:275–80 Eide M, Weinstock M, Clark M (2009) Demographic and socioeconomic predictors of melanoma prognosis in the United States. J Health Care Poor Underserved 20:227–45
Geller A (2009) Educational and screening campaigns to reduce deaths from melanoma. Hematol Oncol Clin North Am 23:515–27 Geller A, Miller D, Swetter SM, Demierre MF, Gilchrest B (2006) A call for the development and implementation of a targeted national melanoma screening program. Arch Dermatol 142:504–7 Geller A, Zhang Z, Sober AJ, Halpern A, Weinstock M, Daniels S et al. (2003) The first 15 years of the American Academy of Dermatology skin cancer screening programs: 1985–1999. J Am Acad Dermatol 48:34–41 Geller J, Swetter SM, Leyson J, Miller D, Brooks K, Geller A (2007) Screening, early detection, and trends for melanoma: current status (2000–2006) and future directions. J Am Acad Dermatol 57:555–72 Heymann WR (2007) Screening for melanoma. J Am Acad Dermatol 56:144–5 Koh H (2007) Melanoma screening: focusing the public health journey. Arch Dermatol 143:101–3
Linos E, Swetter SM, Cockburn MG, Colditz MG, Clarke CA (2009) Increasing burden of melanoma in the United States. J Invest Dermatol 129:1666–74 Lipsker D, Engel F, Cribier B, Velten M, Hedelin G (2007) Trends in melanoma epidemiology suggest three different types of melanoma. Br J Dermatol 157:338–43 Mackie R, Bray C, Hole D, Morris A, Nicolson M, Evans A et al. (2002) Incidence of and survival from malignant melanoma in Scotland: an epidemiologic study. Lancet 360:587–91 Merlino L, Sullivan K, Whitaker D, Lynch C (1997) The independent pathology laboratory as a reporting source for cutaneous melanoma incidence in Iowa, 1977–1994. J Am Acad Dermatol 37:578–85 Moore M, Geller A, Zhang Z, Hayes B, Bergstrom K, Graves J et al. (2006) Skin cancer examination teaching in US medical education. Arch Dermatol 142:439–44 Murray CS, Stockton DL, Doherty VR (2005) Thick melanoma: the challenge persists. Br J Dermatol 152:104–9 Santmyire B, Feldman S, Fleischer A (2001) Lifestyle high-risk behaviors and demographics may predict the level of participation in sun-protection behaviors and skin cancer primary prevention in the United States: results of the 1998 National Health Interview Survey. Cancer 92:1315–24 Schneider J, Moore D, Mendelsohn M (2008) Screening program reduced melanoma mortality at the Lawrence Livermore National Laboratory, 1984 to 1996. J Am Acad Dermatol 58:741–9 SEER (2008) Surveillance, Epidemiology, and End Results Program. Available at www.seer.cancer.gov. Limited-use data (1973–2006), National Cancer Institute, DCCPS, Surveillance Research Program, Cancer Statistics Branch, released April 2009, based on the November 2008 submission. Tejera-Vaquerizo A, Mendiola-Fernandez M, Fernandez-Orland A, HerreraCeballos E (2008) Thick melanoma: the problem continues. J Eur Acad Dermatol Venereol 22:575–9 Weinstock M, Martin R, Risica P, Bwerwick M, Lasater T, Rakowski W et al. (1999) Thorough skin examination for the early detection of melanoma. Am J Prev Med 17:169–75 Weinstock M, Risica P, Martin R, Rakowski W, Smith K, Berwick M et al. (2004) Reliability of assessment and circumstances of performance of thorough skin self-examination for the early detection of melanoma in the Check-It-Out Project. Prev Med 38:761–5
www.jidonline.org
797