The first 2 authors contribute equally to the study.
The incidence and mortality of cutaneous melanoma in southern Germany
Trends by anatomic site and pathologic characteristics, 1976 to 2003
Article first published online: 14 AUG 2006
Copyright © 2006 American Cancer Society
Volume 107, Issue 6, pages 1331–1339, 15 September 2006
How to Cite
Lasithiotakis, K. G., Leiter, U., Gorkievicz, R., Eigentler, T., Breuninger, H., Metzler, G., Strobel, W. and Garbe, C. (2006), The incidence and mortality of cutaneous melanoma in southern Germany. Cancer, 107: 1331–1339. doi: 10.1002/cncr.22126
- Issue published online: 1 SEP 2006
- Article first published online: 14 AUG 2006
- Manuscript Accepted: 31 MAY 2006
- Manuscript Revised: 30 MAY 2006
- Manuscript Received: 17 FEB 2006
- cutaneous melanoma;
- clinical characteristics;
- Central Baden-Wurttemberg;
- Southern Germany
Cutaneous melanoma (CM) incidence and mortality have risen dramatically during the past 2 generations, particularly among Caucasian populations. Detailed, long-term trends of CM in relation to clinical and pathologic characteristics in a Central European population have not been published to date.
The current study was based on 1980 patients with invasive CM diagnosed in Southern Germany during the years from 1976 to 2003 documented by the Central Malignant Melanoma Registry. The German standard population was used to calculate age-standardized rates, and the annual percent change was estimated by using age, anatomic site, histologic type, and tumor thickness.
During the study period, the incidence of CM approximately was tripled for males and females, reaching 10.3 and 13.3 per 100,000 per year, respectively (P < .001). The largest increases occurred for melanoma localized on the upper limbs (annual change, 5.9% for males and 5.0% for females; P < .001) and superficially spreading melanoma (annual change, 7.8% for males and 5.9% for females; P < .001). Thin tumors (Breslow thickness <1 mm) were presented significantly more often during the study period (annual change, 9.8% for males and 6.1% for females; P < .001), predominantly in younger patients. Thick tumors and nodular melanomas were more frequent among older patients (age >70 years), particularly among males. The age-standardized mortality decreased from 1.5 to 0.8 per 100,000 males and from 2.6 to 0.8 per 100,000 females with a significant downward trend for the female population (P < .001).
The current results indicated which diverging trends between incidence and mortality may be explained by improved public awareness regarding suspicious pigmented lesions and the earlier detection of these tumors. Continuation of the current preventive strategy and its expansion to include older age groups in the population are warranted. Cancer 2006. © 2006 American Cancer Society.
Cutaneous melanoma (CM) incidence and mortality have risen dramatically during the past 2 generations in many parts of the world, particularly among Caucasian populations.1–4 Despite the continuous increase in its incidence, recent stabilization or even decline in the mortality from CM has been reported in populations at medium to high risk for this tumor.5–7 The Central Malignant Melanoma Registry (CMMR) of the German Dermatological Society has monitored the incidence and mortality of CM in Southern Germany since the 1970s, when the first evidence of rising trends in Europe emerged in Scandinavia. The main objective of the CMMR is to collect detailed clinical, pathologic, treatment, and follow up data on patients with CM.8, 9 For the area of Central Baden- Wurttemberg (CBW) of Southern Germany, the CMMR collects data not only from hospitals but also from outpatient sources of CM diagnosis. Therefore, the CMMR overcomes the limitation of the nonpopulation-based registries in reporting reliable incidence trends in CM, that is, the under registration of early, curable tumors that are diagnosed and treated on an outpatient basis. Furthermore, as a specialized hospital-based registry, the CMMR collects more clinicopathologic features than normally are recorded by population-based cancer registries and ensures high completeness of registration for various CM characteristics. Currently, the CMMR data base comprises comprehensive information for more than 70,000 patients with CM, which makes it one of the largest CM data bases in the world. Herein, for the first time to our knowledge, we report trends in the incidence and mortality of melanoma in CBW of South Germany during the years 1976 to 2003.
MATERIALS AND METHODS
The CMMR data base was used to identify patients with invasive CM who were diagnosed in the area of CBW in Southern Germany (the counties of Boeblingen, Bahlingen, Reutlingen, and Tuebingen), which has an indigenous population of 923,000 inhabitants according to the Statistic Landesamt (2003 Census). The CMMR is a hospital-based registry. In the CBW area, details on patients with CM are collected primarily from the Hospital of Tuebingen, which is the only university hospital in CBW, and also from private dermatologists and clinics. In CBW, the majority of patients with invasive CM are referred to the Hospital of Tuebingen for treatment, and the cooperative management of patients between dermatologist in private practice and hospital is conducted. Thus, most patients are treated by specialists both in the hospital and on an outpatient basis. After obtaining signed informed consent, the participating physicians complete a registration form at the time patients are diagnosed with CM and forward the form to the CMMR. The participation of private dermatologists in CM registration in CBW regularly is monitored directly by the CMMR. Private dermatologists in CBW are asked to complete a form indicating the patients with CM who were not referred to the hospital for further management of their disease. Mortality data were obtained by cross-checking detailed hospital information obtained by the CMMR with that recorded on death certificates registered with the Statistik Landesamt Baden Wurttemberg. No patients with only death certificate information were included in the study.
The current study was based on 1980 patients who had histologically confirmed, primary CM documented between January 1, 1976 and December 31, 2003. Multiple lesions were excluded. Tumor characteristics and case histories were recorded in a standardized manner and included age, gender, place of birth and residence, tumor thickness, histologic subtype, anatomic site of the tumor, stage of CM according to the American Joint Committee on Cancer (AJCC 2002) staging system, and the date of diagnosis. According the CMMR definition for anatomic sites, hips and shoulders were considered parts of the trunk.
Incidence rates were standardized directly according to the German standard population (age-standardized rates) and to the world standard population (world-standardized rates). Incidence rates were modeled against the year of diagnosis by using regression techniques and were tested for significance by using analysis of linear trend. Quadratic effects were not examined. The estimated annual percentage change (EAPC) in age-standardized incidence rates (% change per year) and its 95% confidence interval (95% CI) were calculated by fitting regression lines to the natural logarithm of age-adjusted rates using calendar year as the regression variable: i.e., Y = bx + c, where Y is the ln (rate), and x is the calendar year, and EAPC = 100*(eb − 1). Testing to determine that the EAPC is 0% is equivalent to testing the null hypothesis that the slope of the line in the above equation is equal to 0. This was tested by comparing b/SE(b) with a t distribution on k − 2 degrees of freedom, where k is the length of the period. The standard error (SE) of b, SE(b), was obtained from the fit of the regression line. The calculation assumed that the logarithm of the rates changed at a constant rate over the entire period. Comparisons of variable distributions between groups were performed by using the chi-square test. Two-sided P values <.05 were considered statistically significant. We used Jointpoint regression analysis with permutation tests to identify significant changes in the trends of age-standardized incidence rates.10 All statistical analyses were performed with the Statistical Package for Social Sciences (SPSS) version 11.5.
From 1976 to 2003, the CMMR recorded 1980 patients with CM (855 males and 1125 females) of German origin in CBW. The participation of private dermatologists and clinics in the registration was >90% during the study period. Ten percent of patients who had CM diagnosed in private clinics were not referred to the CMMR. However, 50% of those CMs were noninvasive, in situ lesions, which were not included in the current study. In Table 1, demographics and clinicopathologic characteristics are presented for 2 consecutive, 14-year periods (1976–1989 and 1990–2003). Between 1976 and 2003, the percentage of male patients increased significantly, and the male-to-female ratio increased from 0.6 to 0.8 (P < .05). A statistically significant increase in the median age at the time of diagnosis also was observed for both genders during the study period. The most prominent anatomic sites for CM were the trunk for men and the lower limbs for women. Throughout the study period, the proportion of CMs of the trunk was decreased significantly in males and corresponded to a slight increase in CMs of the lower limbs (P < .05). Superficial spreading melanoma was the most common histologic type of CM for both genders followed by nodular melanoma and lentigo maligna melanoma. An increase in the proportion of patients with superficial spreading melanoma could be observed along with a distinct decrease in the nodular and lentigo maligna type (P < .01). The percentage of acral lentiginous melanomas increased only in the male population from 1.2% to 3.6%. In CBW, the median tumor thickness decreased steadily from 2mm in males and 1.3 mm in females in 1976 to 0.7mm in males and 0.6 mm in females in 2003 (P < .01) (data not shown). During the study period, the vast majority of invasive CMs measured <1.0 mm in thickness, and only a small percentage of patients presented with tumors that measured >4.00 mm in thickness. Furthermore, the proportion of thin CMs (<1.00 mm) increased from 47.1% and 52.6% to 61.2% and 67.8% in males and females, respectively, coinciding with a decrease of the proportion of thicker tumors (P < .01). A trend toward earlier AJCC stage melanomas also was observed (P < .01) (Table 1).
|Characteristic||No. of patients (%)||P*|
|≤35 y||37 (14.7)||83 (20.7)||74 (12.3)||161 (22.2)|
|36–50 y||81 (32.3)||120 (29.9)||136 (22.5)||160 (22.1)||<.05†,‡|
|51–65 y||78 (31.1)||110 (27.4)||231 (38.2)||208 (28.7)|
|>65 y||55 (21.9)||88 (21.9)||163 (27.0)||195 (26.9)|
|Median age, y||51.3||49.8||57.3||54.0|
|Head/neck||47 (18.7)||63 (15.8)||109 (18.2)||113 (15.7)|
|Trunk||134 (53.4)||94 (23.6)||286 (47.7)||170 (23.6)|
|Upper limb||25 (10.0)||77 (19.3)||70 (11.7)||125 (17.3)||<.05†,‡|
|Lower limb||41 (16.3)||161 (40.4)||108 (18.0)||288 (39.9)|
|Other||4 (1.6)||4 (1.0)||26 (4.3)||25 (3.5)|
|SSM||149 (59.4)||240 (59.9)||409 (67.7)||490 (67.7)|
|NM||50 (19.9)||70 (17.5)||71 (11.8)||69 (9.5)|
|LMM||32 (12.7)||60 (15.0)||40 (6.6)||72 (9.9)||<.01‡|
|ALM||3 (1.2)||11 (2.7)||22 (3.6)||21 (2.9)|
|Other||17 (6.8)||20 (5.0)||62 (10.3)||72 (9.9)|
|Tumor thickness, mm‖|
|<1.00||105 (47.1)||185 (52.6)||339 (61.2)||455 (67.8)||<.01‡|
|1.01–2.00||49 (22.0)||84 (23.9)||105 (19.0)||116 (17.3)|
|2.01–4.00||46 (20.6)||63 (17.9)||73 (13.2)||70 (10.4)|
|≥4.00||23 (10.3)||20 (5.7)||37 (6.7)||30 (4.5)|
|Stage IA||105 (45.5)||181 (50.0)||278 (47.8)||402 (58.2)|
|Stage IB||44 (19.0)||82 (22.7)||138 (23.7)||139 (20.1)|
|Stage IIA||43 (18.6)||58 (16.0)||51 (8.8)||56 (8.1)|
|Stage IIB||14 (6.1)||20 (5.5)||32 (5.5)||27 (3.9)|
|Stage IIC||2 (0.9)||0||8 (1.4)||7 (1.0)||<.01‡|
|Stage IIIA||0||0||7 (1.2)||9 (1.3)|
|Stage IIIB||16 (6.9)||18 (5.0)||46 (7.9)||38 (5.5)|
|Stage IIIC||2 (0.9)||0||6 (1.0)||2 (0.3)|
|Stage IV||5 (2.2)||3 (0.8)||16 (2.7)||11 (1.6)|
The annual incidence of CM rose in both males and females over the 28 years of data collection. In 1976, the age-standardized incidence in CBW was 3.4 per 100,000 per year in men and 4.3 per 100,000 per year in women, and it rose to 10.3 and 13.3 per 100,000 per year in 2003, respectively (Fig. 1A,B). The corresponding EAPC was 5.1% (95% CI, 3.7, 6.4) for males and 4.1% (95% CI, 3.1, 5.1) for females (P < .001). The world-standardized incidence rose from 2.5 to 7.8 per 100,000 per year in men and from 3.1 to 9.3 per 100,000 per year in women over the same period (Fig. 1C,D). The age-standardized mortality decreased from 1.5 and 2.6 per 100,000 per year in 1976 to 0.8 and 0.8 per 100,000 per year in 2003 for males and females. respectively (Fig. 1A,B). The corresponding EAPC of the mortality rates was − 4.9% for females (95% CI, − 7.4 to − 3.9; P = .001) and − 2.0% for males (95% CI, − 4.9 to 0.8; P = .15). The world-standardized mortality decreased from 1.0 to 0.5 per 100,000 per year in men and from 1.3 to 0.4 per 100,000 per year in women over the same period (Fig. 1C,D).
The trends in the age-standardized incidence according to clinicopathologic variables and the corresponding EAPCs are presented in Figure 2 and Table 2, respectively. The age-standardized rates increased for all anatomic sites in both males and females. The largest increases occurred for the upper limbs (EAPC, 5.9% for males and 5.0% for females; P < .001). The incidence of superficially spreading CM increased steeply during the study period (EAPC, 7.8% for males and 5.9% for females; P < .001), whereas there was no consistent trend for the nodular melanoma and lentigo maligna melanoma. The rates of tumors that measured <2 mm in thickness increased significantly during the study period for both males and females, whereas the rates of thicker tumors remained stable. The greatest increase in rates was observed in tumors that measured <1 mm in both males and females (EAPC, 9.8% for males and 6.1% for females; P < .001). Jointpoint regression analysis identified no statistically significant jointpoints, i.e., the best statistical description of the trends was a constant increase at the same rate over the entire study period.
|Characteristic||EAPC (95% CI)*|
|<40 y||4.9 (2.2–7.7)†||4.7 (2.7–6.4)†|
|40–60 y||5.3 (3.3–7.4)†||3.1 (1.7–4.5)†|
|>60 y||6.0 (3.7–8.3)†||4.7 (3.0–6.3)†|
|Head/neck||4.3 (2.1–6.5)†||2.8 (0–5.8)‡|
|Trunk||5.2 (3.4–7.0)†||3.2 (1.5–5.0)†|
|Leg||5.3 (2.8–7.8)†||4.2 (2.6–5.8)†|
|Arm||5.9 (3.7–8.2)†||5.0 (2.2–7.9)†|
|SSM||7.8 (5.7–10.1)†||5.9 (3.6–8.0)†|
|NM||1.5 (−0.6–3.6)§||5.9 (−2.9–1.8)§|
|LMM||−1.2 (−3.6–1.3)§||0.0 (−2.4–2.4)§|
|Tumor thickness, mm|
|<1.00||9.8 (7.0–12.7)†||6.1 (4.0–7.8)†|
|1.01–2.00||5.3 (2.5–8.2)†||3.8 (0.5–7.1)‡|
|2.01–4.00||1.4 (−1.6–4.5)§||1.6 (−1.6–3.8)§|
|≥4.00||−0.6 (−4.4–3.5)§||0.5 (−2.2–3.1)§|
The distribution according to anatomic site, histogenetic type of CM, and tumor thickness (<1 mm, 1–2 mm, 2–4 mm, and >4 mm) for 7 age groups (age <30 years, ages 31–40 years, ages 41–50 years, ages 51–60 years, ages 61–70 years, ages 71–80 years, and age > 80 years) is shown in Figure 3. Thin tumors (<1 mm) and superficially spreading melanomas were more frequent at younger ages in both males and females. Thicker tumors (>2 mm) and nodular melanomas were more common in older age groups (age > 70 years), particularly among males. The proportion of CMs of the head and neck and of the lentigo maligna melanoma type increased with age among both genders.
Based on data from the CMMR of the German Dermatological Society during a 28-year period, this report analyzes the incidence of and mortality from CMs diagnosed in 1980 patients of German origin. To our knowledge, this is the first report of detailed, long-term trends in the incidence of CM by clinical and pathologic characteristics in a Central European population. We have demonstrated a persistent increase in the incidence of CM presenting in Southern Germany from 1976 to 2003. The age-standardized rates tripled during this period for both genders, reaching 10.3 and 13.3 per 100,000 inhabitants per year for males and females, respectively. In females, the incidence was approximately 30% higher than in males, and this difference persisted throughout the study period. The mortality from CM was reduced in both males and females, with trends reaching statistical significance only in the female population.
The CMMR works as a hospital-based registry, and the data are recorded with the informed consent of the patients. The recording of data on patients with melanoma is implemented into the clinical practice, which has 2 major advantages concerning data quality: First, the surveillance and follow-up examinations are based on this data set, and this guarantees a continuous check of the accuracy of the basic tumor data. Second, follow-up data on the patients are generated regularly by the routine clinical surveillance procedures. Furthermore, in the University Department of Dermatology in Tuebingen, all external histopathologic diagnoses are scrutinized by internal reference histopathology with a standardized reporting schedule, and strict follow-up procedures are practiced, including a reminder system if patients do not appear at the appointed date. Thus, the accuracy and completeness of the data generated here may be better compared with that generated by normal, population-based cancer registries.
The incidence rates observed in this study were comparable to those reported by the Cancer Registry of Saarland, which is currently the only population-based registry in Germany with data quality that is comparable to that of international cancer registries.11, 12 Similar rates also are reported by population-based registries of neighboring France, Switzerland, and Italy, which report world-standardized rates of approximately 10 per 100,000 population per year.11, 13 This is more likely to be the representative incidence of CM in Central Europe. The mortality-to-incidence ratio in the current study was comparable to that reported from high-quality registries in Europe (i.e. Scottish, Swedish) for the corresponding periods.6, 11
We demonstrated that the incidence of CM in Southern Germany was increasing, especially in thin lesions with a good prognosis. The median tumor thickness of CM diagnosed between 1976 and 2003 decreased significantly in both genders, reaching 0.7 mm, which is comparable to that reported for high-incidence regions of Australia, where there have been well organized preventive interventions for several years.13 In agreement with the increase in the rates of thin tumors and the stabilization of thicker, untreatable lesions is the observed leveling off in mortality from CM noticed in the current study. It is noteworthy that, among females, in whom the trends toward thinner lesions were more prominent, the reduction in mortality also was higher than among males and reached statistical significance. Of particular interest is the finding that the decrease in the mortality rates was not coupled with a decrease in the thicker, untreatable tumors. It seems that, as the incidence is truly increasing, additional patients also are diagnosed with thicker tumors, supporting their lack of reduction.
The unexpected combination of stable rates of the thicker tumors and decreasing mortality rates may imply a better outcome for patients who have thicker CMs through the years. This can be investigated with detailed survival analyses of our population, which is the objective of our next studies. Nevertheless, except for secondary prevention, a number of new modalities in the management of CM were established during the last 3 decades (such as routine excision with safety margins) along with standardized surveillance programs, which enable the early detection of recurrences.14 In addition, new diagnostic procedures that enable earlier detection of first recurrences emerged, particularly lymph node ultrasound.15 However, the degree to which all of these innovations may contribute to the observed decrease in mortality rates remains unclear.
Analyzing the characteristics of patients who present with advanced tumors reveals specific groups of the population that do not seem to benefit from current interventions and that should be the focus of our preventive strategies.9 We showed that these groups may include patients age >70 years, particularly males who present with nodular melanomas. This finding is in agreement with recent reports from the U.S. and Australia.16–19 Based on national and regional data, the rates of nodular melanoma among whites in the U.S. decreased significantly during the last decade in individuals age <45 years, whereas an increase was observed among individuals ages ≥65 years. The median thickness of nodular melanomas diagnosed from 1995 to 2000 was 3.0 mm, whereas the median thickness of superficially spreading tumors was only 0.55 mm.18 In Australia, the incidence of thick melanoma (≥1 mm) increased significantly only in the male population during 1992 to 2002; however, compared with thinner lesions, this increase clearly was lower.17 In the same population, nodular type and older age were associated independently with thick, poor-prognosis tumors.20 It appears that nodular melanomas represent a significant obstacle to the reduction of mortality. The asymmetry, border irregularity, color variegation, and diameter <6 mm (ABCD) system for the recognition of early melanoma does not apply well to this histologic type, and the shorter period of growth limits the opportunity for early detection, even in patients under regular surveillance for melanoma screening.21 There is accumulating evidence that the addition of an “E” for “evolving” to the ABCD system may improve and enhance early recognition by clinicians and laypersons.22 We believe that patients and their immediate environment should bear the major responsibility for the early detection of melanoma. However, the early features of nodular melanoma need to be investigated and clarified so that they can be promoted to the population and to primary care professionals.19–23
This focus should not be at the expense of the general preventive educational reminders. In Germany, skin cancer screening is recommended in the frame of the preventive cancer screenings for breast cancer and prostate cancer (Gesetzliche Krebsfruherkennungs-massnahmen, section 181 RVO) that have been reimbursed by the German health care system starting for men at age 45 years and for women at age 30 years since 1975. However, skin cancer screening is not reimbursed separately.12, 24 To our knowledge, statutory general skin cancer screening programs do not exist in other European countries, in the U.S., or even in Australia. This is mainly because of the low overall mortality from skin cancer and because there are not data from randomized, controlled trials to demonstrate the benefit of skin cancer screening in the general population.25 In any case, skin cancer screening remains appealing for this common and visible cancer, especially for melanoma, which is eminently treatable when detected early. Physician-detected melanomas usually are thinner than self-detected or family-detected melanomas, and it is well known that screening also provides an opportunity for education and advocacy.25 Indeed, during the last 30 years, a great effort has been made in the area of CBW to increase public and professional awareness of suspicious pigmented lesions. Most prevention programs have combined messages of solar protection with those of early detection, and there is some epidemiologic evidence that these interventions can achieve changes associated with reduced mortality.4 If current preventive strategies continue and are reinforced to attend specific groups of the population, as discussed above, we could expect further reduction in the mortality from CM the coming years. However, the incidence of CM in Southern Germany probably will keep increasing during the next years, and recent estimates suggest a doubling of rates until 2015.26
The factors underlying the rapid increase in the incidence of CM are understood incompletely, but increased total exposure to the sun and altered patterns of exposure are implicated strongly.27 Changes in life style in affluent Western societies that have lead to increased leisure time spent outdoors, especially during holidays in sunny climates, have been proposed as most likely responsible for the observed increases in incidence rates of CM in these populations.28 Moreover, in the last 20 years, sun-bed use has exploded in light-skinned populations, mainly among young female adults, and a recent meta-analysis of case–control studies indicated a significant increased risk of CM subsequent to sun-bed/sun-lamp exposure.29 In the current study, a distinctive pattern between the sites most likely and least likely to be exposed intermittently to the sun was not clear based on the estimated annual change in the rates.27 However, superficially spreading melanomas, which have been associated with intermittent sun exposure, increased more steeply during the study period, whereas lentigo maligna melanomas, which are associated with heavy, chronic sun exposure, did not show a significant trend.30, 31
In summary, the age-standardized incidence rates of CM in Southern Germany during the years 1976 to 2003 almost tripled, reaching 10.3 and 13.3 per 100 000 per year for males and females, respectively. The increase was significantly greater at each anatomic site for males. Tumors with a thickness <1 mm presented significantly more often through the years and were predominant in younger age groups. The rates of thicker tumors (>2 mm) and nodular melanomas remained stable and were more frequent among older age groups (age >70 years), particularly in males. The median tumor thickness was reduced to 0.7 mm in both genders, revealing a significant trend toward prognostically more favorable CMs. Correspondingly, the mortality rate from CM according to the German standard population was reduced in both males and females, with a significant downward trend for the female population. The diverging trends between incidence and mortality may be explained by improved public awareness regarding suspicious pigmented lesions and the earlier detection of this tumor. Continuation of the current preventive strategy and its expansion to include older age groups is warranted.
- 1Changing epidemiology of malignant cutaneous melanoma in Europe 1953–1997: rising trends in incidence and mortality but recent stabilizations in western Europe and decreases in Scandinavia. Int J Cancer. 2003; 107: 119–126., , , .
- 11World Health Organization. GLOBOCAN: Cancer Incidence, Mortality and Prevalence Worldwide. IARC cancer base no. 5, version 1.0. Lyon: IARC; 2001.
- 15Ultrasound examination of regional lymph nodes significantly improves early detection of locoregional metastases during the follow-up of patients with cutaneous melanoma: results of a prospective study of 1288 patients. Cancer. 2000; 88: 2534–2539., , , , , .