Fax: (011) 0049-07071-295187
Age and gender are significant independent predictors of survival in primary cutaneous melanoma
Article first published online: 27 FEB 2008
Copyright © 2008 American Cancer Society
Volume 112, Issue 8, pages 1795–1804, 15 April 2008
How to Cite
Lasithiotakis, K., Leiter, U., Meier, F., Eigentler, T., Metzler, G., Moehrle, M., Breuninger, H. and Garbe, C. (2008), Age and gender are significant independent predictors of survival in primary cutaneous melanoma. Cancer, 112: 1795–1804. doi: 10.1002/cncr.23359
- Issue published online: 4 APR 2008
- Article first published online: 27 FEB 2008
- Manuscript Accepted: 30 OCT 2007
- Manuscript Revised: 27 OCT 2007
- Manuscript Received: 20 JUN 2007
- cutaneous melanoma;
- prognostic factors;
- clinical presentation
The aim was to identify age- related and gender-related differences in the clinical presentation and outcome of patients with primary cutaneous melanoma (CM).
A total of 4785 CM patients without clinical evidence of metastasis diagnosed during the period 1976–2001 in southern Germany included in the analysis. Kaplan-Meier analyses were performed to estimate and to compare disease-specific survival (DSS) and survival after first recurrence (SAR). The Cox proportional hazards model was used to evaluate the effect of multiple variables on DSS and SAR.
Increasing age and male gender were independently associated with thicker tumors (>2 mm) and histologic ulceration (P < .001). Patients older than 65 years had lower 10-year DSS than younger patients (81.8% vs 88.4%, P < .001) and this difference was more pronounced in women (P < .001) than in men (P = .06). Males had lower 10-year DSS than females (83.5% vs 88.5%, P < .001) but this difference did not reach statistical significance in patients older than 65 years (P = .162). In multivariate analysis adjusted for tumor thickness, ulceration, anatomic site, histologic subtype, DSS, site of first recurrence, time trend, sentinel lymph node status, age, and gender were independent predictors of DSS and SAR (P < .05).
Older age and male gender are associated with prognostically unfavorable primary CM. Expansion of current preventive strategies to target these subgroups is warranted. Moreover, age and gender are independent predictors of the outcome of CM patients. Females have a better prognosis than males but this difference disappears after the age of 65. Younger patients have a more favorable prognosis than older patients, a difference more pronounced in women. Cancer 2008. © 2008 American Cancer Society.
The incidence of cutaneous melanoma (CM) has been increasing during the last 30 years in most Caucasian populations.1–3 Since the mid-1980s, however, the rates of increase began decreasing or at least leveling off.4–9 Recent studies showed that the observed leveling-off in rates predominantly occurred in younger persons and the patterns showed positive effects for cohorts born after 1945 or 1950.4, 6, 9 Conversely, the rate of increase in melanoma incidence and mortality has been significantly higher for age groups older than 60 years in Europe, the US, and Australia.1–4 Older persons present more frequently with advanced tumors, lentigo malignant melanomas (LMM), and melanomas arising in sun-exposed anatomic areas and the disease-specific survival (DSS) of melanoma patients over 60 years old is significantly shorter compared with younger ages.10 Thus, age seems to be a significant prognostic factor for melanoma and its prognostic significance has been demonstrated likewise after adjusting for all known prognostic variables such as tumor thickness, ulceration, etc.11 This may suggest that the role of age in the natural course of melanoma might not be entirely explained by differences in the proportions of the known clinicopathologic variables of this disease.
Significant differences also exist in melanoma incidence and mortality between males and females. In most Caucasian populations the incidence rates for melanoma are higher for women compared with men, but the rate of increase during the last 3 generations has been significantly higher for males.1, 2, 4 Moreover, women present more frequently with early diagnosed tumors and experience longer survival compared with men. As a probable result, the mortality from melanoma in women was reduced in many medium- to high-incidence areas.1, 2, 4 No reduction in the mortality from melanoma in males has been reported so far.
Herein, we aimed to identify age- related and gender-related differences in the clinical presentation and outcome of patients with CM.
MATERIALS AND METHODS
We included in the analysis 4785 patients diagnosed with primary CM during the period 1976–2001 at the University Medical Center of Tuebingen in southern Germany. Only patients diagnosed with histopathologically confirmed primary invasive (Clark level >I) CM without clinical evidence of metastasis after all investigations performed at the time of primary diagnosis were included in the analysis. Ocular and mucosal melanomas were excluded. Patients with positive sentinel lymph nodes were also excluded because this information was not available for the early years of the study. In our sample, 602 (12.6%) patients were subjected to sentinel lymph node biopsy (SLNB) and 72 (1.2%) were found to have lymph node micrometastasis. The remaining 4183 (87.4%) had only clinical staging at onset of diagnosis. Multiple lesions and patients younger than 14 years old were excluded. Data were retrieved from the electronic database of the Central Malignant Melanoma Registry (CMMR), which is a hospital-based registry. Written informed consent was obtained from all patients in the current study. Data obtained for each patient included age, gender, place of birth, date of diagnosis, anatomic localization and histopathologic characteristics of primary melanoma, date of diagnosis of the first recurrence, and site of first recurrence. Site of recurrence was categorized into 3 groups according to the definitions of Gershenwald et al.12: 1) local or satellite or in transit recurrence; 2) metastasis to the regional lymph nodes; or 3) metastasis to distant anatomical sites. No data regarding the number of affected regional lymph nodes were available. Aims and methods of data collection by the CMMR have been previously reported in detail.2, 13
Comparisons of the distribution of categorical variables between groups were performed using the Pearson chi-square test. Multivariate logistic regression analysis was used to adjust for the effects of multiple variables. For the comparisons of the distribution of continuous variables, Wilcoxon rank-sum tests and t-tests were used. Follow-up time was defined as the date of last follow-up or death minus date of operation (DSS) or the date of diagnosis of first recurrence (survival after recurrence, SAR). Disease-free survival (DFS) was defined as the date of diagnosis of first recurrence minus the date of operation. Follow-up time was described as a median value with interquartile range (IQR). Follow-up time was cut to a maximum of 10 years because patients usually participate in the CM follow-up program for a maximum of 10 years, whereas afterward only patients with progressive disease are regularly documented, and little information can be retrieved for patients without recurrences. Only deaths caused by CM were considered in survival analysis. All 4785 CM patients (clinically and pathologically staged) were included in survival analysis. Melanoma-specific survival curves and estimated survival rates with relative 95% confidence intervals (95% CIs) were generated according to the Kaplan-Meier method and were compared by the 2-sided log-rank test. Age was classified in 2 subgroups (<65 years, ≥65 years). An age threshold of 65 years to define “elderly” was chosen because it is regarded as the normal retirement age in many Western industrialized countries. Multivariate survival analysis was performed using Cox modeling. In the multivariate analyses age and tumor thickness were entered as continuous variables. Categorical variables were dummy coded. The Cox model was described by means of hazard ratios (HRs) together with 95% CIs and P-values were based on the Wald test. Examination of the graphic plot of the log-log survivor function for the variables entered in the analysis showed no violation of the proportional hazards assumption. Throughout the analysis 2-sided P-values less than .05 were considered statistically significant. All statistical analyses were performed with the Statistical Package for Social Sciences (SPSS) v. 15.0.1 (SPSS, Chicago, Ill).
The distribution of various clinical and pathologic characteristics of melanoma patients stratified by age at the time of primary diagnosis is presented in Table 1. Overall there were 3665 (76.6%) patients younger than 65 years and 1120 (23.4%) patients 65 years or older. In both younger and older age groups no significant differences with respect to gender were observed (P = .169). Older patients of both genders were more likely to have head and neck melanomas (30.5% vs 9.5%, P < .001), which was mainly because of a 3–5-fold increase in the proportion of tumors arising on the face. The proportion of tumors arising on the scalp or neck increased less markedly with age. On the other hand, the proportion of tumors of the back (19.0% vs 27.9%, P < .001) or buttocks (1.6% vs 4.2%, P < .001) decreased significantly in older patients of both genders. The most common histologic type of melanoma was superficially spreading melanoma (SSM) in both age groups (77.1% in patients younger than 65 years vs 52.0% in patients 65 years or older, P < .001). The proportion of LMM (21.2% vs 4.3%, P < .001), acral lentiginous melanoma (ALM) (6.1% vs 1.9%, P < .001), and nodular melanoma (NM) (18.1% vs 13.0%, P < .001) was significantly higher in patients 65 years or older. Older patients presented more frequently with advanced, prognostically unfavorable tumors. The proportion of thin (<1 mm) lesions was significantly lower in ages 65 years or older (46.9% vs 63.3%, P < .001). Likewise, older patients presented more frequently with ulcerated tumors (9.9% vs 4.7%, P < .001).
|<65 Years||≥65 Years||P*|
|% Males n = 1586||% Females n = 2079||% Males n = 511||% Females n = 609|
|Face||4.8||4.6||17.0||26.9||<.001 between ages and sexes|
|Skin of genital area||0.1||0.2||0.2||1.1|
|Upper upper limb||7.5||11.4||6.3||12.6|
|Lower upper limb||4.0||5.0||3.1||5.6|
|Upper lower limb||8.0||14.3||2.7||3.6|
|Lower lower limb||6.7||24.8||4.5||24.1|
|SSM||75.7||78.2||55.8||48.8||NS between genders <65 y|
|NM||13.7||12.4||19.2||17.2||.022 between genders ≥65 y|
|LMM||4.2||4.3||17.0||24.6||<.001 between ages|
|Level of invasion|
|II||20.1||23.9||14.3||18.6||<.001 between genders <65 y|
|III||46.2||48.1||34.6||32.2||NS between genders ≥65 y|
|IV||28.6||24.3||40.7||41.1||<.001 between ages|
|≤1 mm||59.8||65.9||45.0||48.4||<.001 between genders <65 y|
|1.01–2.00 mm||21.9||21.2||23.9||23.3||NS between genders ≥65 y|
|Median IQR||0.82 (1.10)||0.74 (0.86)||1.10 (1.92)||1.10 (1.75)||†<.001 between genders|
|.094 between ages|
|Yes||6.4||3.3||11.5||8.5||<.001 between ages, genders|
|No||93.6||96.7||88.5||91.5||<.001 between genders <65 y|
|.104 between genders ≥65 y|
In Table 2 the distribution of site of first metastasis of all 4785 patients (clinically and pathologically staged) is presented. There were 865 patients who developed metastasis. Of them, 128 (16.7%) had previously undergone SLNB and 45 (5.4%) had lymph node micrometastasis. The proportion of patients <65 years who developed metastasis was significantly lower compared with older ages (16.3% vs 21.1%, P < .05). The pattern of first metastasis was significantly different between age groups. Older patients presented more frequently with local recurrences (34.3% vs 20.0%, P < .001) and less frequently with recurrence in the regional lymph nodes (37.7% vs 50.9%, P < .001) compared with younger ages. However, SLNB was proportionally more common among patients older than 65 years (17.3% vs 12.0%, P < .001) (data not presented). The proportion of patients who presented with distant metastasis as first site of recurrence was not significantly different between the 2 age groups (Table 2). However, older patients presented more frequently with liver metastasis (18.2% vs 7.5%, P < .05) and brain metastasis (24.7% vs 15.2%, P < .05) (data not presented).
|<65 Years||≥65 Years||P*|
|% Males n = 318||% Females n = 281||% Males n = 141||% Females n = 125|
|Site of first metastasis|
|Local/satellite/in transit||15.1||25.6||26.1||41.6||.028 between age groups in males|
|Regional lymph nodes||52.8||48.8||43.2||32.8||.002 between age groups in females|
|Distant||32.1||25.6||30.6||25.6||.004 between genders <65 y|
|.042 between genders ≥65y|
The median follow-up was 7.6 years (IQR, 5.3) in patients <65 years, and 5.0 years (IQR, 4.9) in patients 65 years or older. Ten-year DFS probability was significantly lower in patients 65 years or older (71.9%, 95% CI, 68.4–75.4, P < .001) compared with younger patients (81.8%, 95% CI, 80.4–83.2, P < .001). Ten-year DSS probability was significantly lower in patients 65 years or older (81.8%, 95% CI, 78.5–85.1) compared with younger patients (88.4%, 95% CI, 87.2–89.6, P < .001). As shown in Figure 1A,B, the latter difference was more pronounced in women (P < .001) compared with men, where it was of borderline statistical significance (P = .06). Five-year SAR probabilities were lower in patients 65 years or older who developed their first recurrence in local anatomic sites (59.7 vs 66.6%, P = .3) but this difference did not reach statistical significance (Fig. 1C). Five-year SAR probabilities were significantly lower in patients 65 years or older as compared with younger patients, who developed their first metastasis in regional lymph nodes (33.8% vs 44.3%, P = .031) (Fig. 1D). No difference in terms of SAR was observed between the 2 age groups in patients who developed their first metastasis in distant anatomic sites (P = .380) (data not presented).
The anatomic subsite of tumors differed significantly by gender. In men the trunk was the most common anatomic localization of CM (55.1%), whereas women were more likely to have lower extremity lesions (42.7%) (P = .001). The proportion of LMM was significantly higher in women compared with men in the older (24.6% vs 17.0%, P = .02) but not in the younger age group (P = .453). The proportion of NMs was slightly higher in males compared with females (15.1% vs 13.5%, P = .133). Male patients presented more frequently with advanced, prognostically unfavorable tumors. The median tumor thickness in males 65 years or older was 0.82 mm compared with 0.74 mm in females of the same age group (P < .001). No difference in the median tumor thickness between males and females older than 65 years old was observed (Table 1). Likewise, in the ages 65 years or older, women presented less frequently with ulcerated tumors than men (3.3% vs 6.4%, P < .001). This difference was not statistically significant in the older age group (P = .104).
During follow-up 459 (20.0%) men and 406 (15.1%) women developed metastasis (P < .05) (Table 2). The pattern of first metastasis differed significantly by gender. Women presented more frequently with local recurrences as first site of metastasis (30.5% vs 15.9%, P < .001) and less frequently with regional lymph node (43.8% vs 50.3%, P < .001) or distant recurrence (25.6% vs 35.1%, P < .001). The pattern of distant metastasis did not differ significantly between males and females (P = .394).
Ten-year DFS was significantly lower in men (75.8%, 95% CI, 73.4–78.2) compared with women (82.7%, 95% CI, 80.9–84.5, P < .0001). Ten-year DSS was significantly lower in men (83.5%, 95% CI, 81.5–85.5) compared with women (88.5%, 95% CI, 88.5–91.3, P < .0001). When our analyses were stratified by age group no significant preponderance of women regarding the DFS was observed in patients 65 years or older (P = .485) (data not presented). Likewise, women 65 years or older had a slightly more favorable outcome in terms of DSS compared with men of the same age group, but this difference did not reach statistical significance (P = .162) (Fig. 1A,B).
Five-year SAR probabilities were significantly lower in male patients as compared with females, who developed their first metastasis in local anatomic sites (48.9% vs 72.4%, P = .022) or regional lymph nodes (36.6% vs 49.5%, P = .031) (Fig. 2C,D). Five-year SAR probabilities did not differ significantly between males and females who developed their first recurrence in distant anatomic sites without reaching statistical significance (P = .932) (data not presented).
In the multivariate logistic regression analysis, age (P < .001) and male gender (P = .002) were associated with thicker (>2 mm) tumors. Other factors independently associated with thicker melanomas were the presence of histologic ulceration (P < .001) and the histologic subtype other than SSM (P < .001). Similarly, age (P < .001) and gender (P < .001) were associated with ulcerated tumors. Other factors independently associated with the presence of histologic ulceration were increasing tumor thickness (P < .001), Clark level of invasion (P < .001), histologic type of NM or LMM (P < .001), and anatomic localization other than head and neck (P < .001).
The results of the multivariate Cox regression analysis are presented in Table 3. In the model showing the main effects of the variables (without interaction terms) age (P = .005) and gender (P = .01) were independent predictors of DSS. Other independent determinants of DSS were tumor thickness (P < .001), histologic ulceration (P < .001), Clark level of invasion (P < .001), anatomic site of primary melanoma (P < .001), histologic subtype (P = .006), SLNB positivity (P < .001), and calendar year of diagnosis (P < .001). Inclusion of a term for the interaction of age and gender as suggested by Figures 1A,B and 2A,B resulted in the model presented in Table 3. The age/gender interaction term was a highly significant predictor of DSS, in contrast with age, which was no more significant at the .05 level. The latter suggests that the prognostic effect of age depends on gender and that the impact of age on the outcome of melanoma is stronger in women than in men. This finding is consistent with the results of the univariate analysis presented in Figures 1A,B and 2A,B. Likewise, age (P = .014) and gender (P = .01) were independent predictors of SAR. Other independent predictors of SAR were the DFS (P < .001) and the site of first recurrence (P < .001).
|Main effects without interactions||Adjusted for age-gender interaction|
|HR (95%CI)*||P†||HR (95%CI)||P|
|Age, y||1.010 (1.003–1.017)||.005||0.98 (0.96–1.00)||.095|
|Sex (women vs men)||0.8 (0.6–0.9)||.010||0.3 (0.1–0.6)||.001|
|Age-gender interaction||—||—||1.029 (1.006–1.032)||.005|
|Tumor thickness (mm)||1.20 (1.16–1.25)||<.001||0.94 (0.93–0.96)||<.001|
|Clark (III/IV vs II/III)||2.4 (1.9–3.0)||<.001||2.4 (1.9–3.0)||<.001|
|Ulceration (yes vs no)||2.2 (1.6–2.9)||<.001||2.2 (1.6–2.9)||<.001|
|SLNB (positive vs negative)||2.4 (1.5–3.9)||<.001||2.6 (1.6–4.1)||<.001|
|NM||1.4 (1.1–1.8)||.009||1.4 (1.1–1.8)||.010|
|LMM||0.9 (0.6–1.4)||.670||0.9 (0.6–1.4)||.559|
|ALM||2.0 (1.3–3.1)||.003||1.9 (1.3–3.0)||.003|
|Other||1.0 (0.5–1.8)||.977||1.0 (0.5–1.8)||.961|
|Trunk||1.3 (1.0–1.8)||.084||1.4 (0.9–1.9)||.066|
|Upper limb||0.6 (0.4–0.9)||.009||0.6 (0.4–0.9)||.010|
|Lower limb||0.9 (0.7–1.3)||.703||0.9 (0.7–1.3)||.777|
|Calendar year of diagnosis||0.94 (0.93–0.96)||<.001||1.20 (1.16–1.25)||<.001|
In a hospital-based study of 4785 patients with primary CM without clinical evidence of metastasis at the time of diagnosis we observed significant differences regarding melanoma clinical presentation and natural course between men and women as well as between patients older and younger than 65 years. Women and patients younger than 65 years old had significantly longer survival compared with men and older patients, respectively. Because these conclusions depend critically on adequately controlling for differences in clinical presentation of melanoma, we used stratified and multivariate analyses to adjust for other prognostic factors. Increasing age and male gender were associated with prognostically unfavorable primary melanomas and shorter DSS. These effects were independent from histologic characteristics and anatomic localization of primary tumor.
This study shows that the elderly comprise an important high-risk group among melanoma patients. Older age was significantly associated with all well-established prognostic factors such as tumor thickness, histologic ulceration, and nodular type. Several explanations have been proposed to explain the presentation of elderly patients with thicker melanomas. Older persons usually give less importance to skin changes and they perform self-examination less often compared with younger persons.14 Analysis of the US National Skin Cancer Screening program showed that elderly males represented only 25% of the participants but constituted approximately half of melanoma cases discovered.15 Low melanoma awareness and ignorance of the early signs and symptoms of melanoma has been associated with delay in seeking medical attention for CM, thus providing the lesion the opportunity to progress.14, 16 There is evidence that older patients are less likely to report changes in elevation/color of their lesions, signs that are associated with thin melanomas, but more likely to report ulceration and bleeding, indicating advanced, more aggressive tumors.17
Conversely, older patients may have no greater difficulty in discriminating early changes of pigmented lesions, as has been shown in a study where patients were asked to identify early signs of melanoma in photographs.18 This might suggest that older patients simply do not make use of their skills in detecting melanomas and that public education campaigns might encourage them to do so. In the same study, however, both younger and older patients had poor ability to detect changes of melanoma after longer time periods. The latter might be quite relevant to the late presentation of older persons with CM because of the increased proportion of LMM, the histologic type which develops slowly over a long period of time.
The increased proportion of NMs in the elderly is often presented as a significant obstacle to the early diagnosis of CM. In contrast with LMM, NM develops de novo, rapidly, and lacks, by definition, adjacent radial growth phase components that are associated with early melanoma signs and symptoms.17 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.19 The addition of an “E” for “evolving” to the ABCD system may improve and enhance early recognition by clinicians and laypersons.20 We advocate the expansion of current preventive educational reminders to the elderly as well as to their immediate environment. Investigations aiming to clarify early features of melanoma in the elderly and promotion of these features to the population and to primary care practitioners would be valuable in secondary prevention of melanoma in the older ages.2, 21
Other explanations for the late diagnosis in elderly individuals is deteriorating vision, loss of partner, development or increase of seborrheic keratoses, with which melanoma can be easily confused, small nevi density, which leads to lower consciousness of melanoma risk, and development of a higher proportion of melanomas in hard-to-see anatomic sites (ie, back).22 Unfortunately, data regarding the presence of pigmented lesions other than melanoma were not available for analysis in our study. However, even if the proportion of melanomas developing on the scalp or foot was significantly higher in patients older than 65 years, we could not confirm the increase in the proportion of melanomas arising on the back of elderly persons, as reported by others. Conversely, Hanrahan et al.23 showed that for any given thickness and histologic types of melanoma older individuals are less likely to recognize early melanoma compared with younger individuals. This is in line with our findings that older age is associated with thicker or ulcerated tumors independent of anatomic site and histologic subtype.
We showed that older persons present more frequently with local recurrence (satellite, in transit) and less frequently with regional lymph node recurrence compared with younger patients. This finding is in line with an apparent paradox, in that older patients have a lower sentinel lymph node metastasis rate yet have a decreased DSS compared with younger patients.24, 25 The finding that older age and gender influence SAR independently from site of first recurrence, disease-free interval, and histologic characteristics of primary tumor suggests that the survival benefit of younger patients might not be solely attributed to differences in the presentation of primary tumors. Immunosenescence resulting in decreased capacity of the organism to fight malignancy, changes in host immune biology, and decreased ability to repair DNA damage are frequently cited as possible explanations for the age differences in the survival of cancer patients, but the relation between age and the biology of melanoma is poorly understood.25
Age-related differences in the treatment of older patients might have also contributed to the observed discrepancy in the survival between patients younger and older than 65 years. With increasing age there is accumulation of medical comorbidity and increased physiological frailty that raise the mortality and morbidity related to aggressive surgery, radiotherapy, and chemotherapy.26, 27 Moreover, chronologically advanced age represents a risk factor for undertreatment, even after adjustment for markers of physical frailty, comorbidity, and social support.28 Elderly patients are generally underrepresented in clinical trials and the results from clinical trials in younger patients might not be directly applicable to them for several reasons (pharmacokinetic differences, different side effects, comorbidity, reduced cognitive function, polypharmacy, etc). Exploration of practical means to improve the accrual of older patients into clinical trials could provide better evidence for treatment of the elderly.29 Finally, nonclinical factors may also significantly limit the access of older patients to healthcare. Socioeconomic factors, difficulties in transportation to clinical centers, differences in education, and stronger relationships with primary care physicians present as common barriers to the participation of senior patients in clinical studies and more aggressive treatment strategies.30 Conversely, so far it has not been shown conclusively that patient management has contributed to any survival benefit in patients with progressive disease.2
In our study, female gender was independently associated with longer DSS and SAR. Similar results have been reported by other groups but no clear explanation for this difference exists so far.31 Gender differences in the clinical presentation of melanoma may exert a significant role on the outcome of the disease. Women present more frequently with melanomas in their limbs, whereas men have more trunk lesions, which are associated with shorter survival. The proportion of NM did not differ significantly between males and females in our population, but women presented more frequently with thinner, nonulcerated melanomas. This finding is often explained by higher attention of women to their bodies and greater awareness of healthcare issues, even though a possible role of biologic differences cannot be ruled out.27
The recurrence rates for women were significantly lower than for men and female gender was associated with a higher proportion of local and regional recurrences and a lower proportion of distant recurrences compared with men. The finding that gender was independently associated with longer DSS and SAR points to a more direct role of gender in the natural course of melanoma. Estrogens were initially suggested to play a protective role against melanoma but the preponderance of evidence now is that melanoma is not hormone-dependent. Studies on the influence of pregnancy in melanoma have been inconclusive and clinical trials of antiestrogens have almost uniformly disappointing results.32, 33 Further research is necessary to elucidate the complex mechanisms that underline the gender-related differences in the outcome of melanoma with the hope that the survival advantage of women might be expanded to include the entire melanoma population.
Our study has several limitations. Because our sample is not population-based, the results may not be extrapolated easily to the general population. However, it can be argued that this sample is largely representative of the situation in southern Germany, as regional population-based incidence studies confirm.2 Unfortunately, we had no data regarding the mitotic rate of primary tumor, which was recently shown to be an independent prognostic variable in CM patients.34 In addition, the sentinel lymph node status of our patients diagnosed during the early years of the study was not available, and for this reason we included all patients without clinical evidence of metastasis. Although SLNB significantly prolongs DFS, so far there is no evidence that this procedure has an impact on the DSS of patients with melanoma.35 Third, the number of involved lymph nodes in patients who developed regional lymph node metastasis was not available in our database and this factor is the major determinant of the outcome in this patient subgroup. Finally, we had no details regarding systemic treatments administered to our patients and socioeconomic factors that are known to influence the outcome of cancer patients. However, we had high-quality data with negligible rates of unknown or missing cases and almost complete long-term follow-up data for our entire collective. The variable ulceration of the primary tumor, however, may have been underreported during the early years of the study, because at that time its prognostic significance was not yet established.
In conclusion, in this study we showed significant age- and gender-related differences in the presentation and in the outcome of patients diagnosed with primary CM. Older patients presented more frequently with late-diagnosed, prognostically unfavorable tumors. Therefore, expansion of current preventive strategies to target these subgroups is warranted. Consequently, the DSS of elderly and female patients was generally lower compared with younger and female patients, respectively. However, the preponderance of women in terms of DSS disappears after the age of 65. In addition, the survival difference between older and younger patients is more pronounced in women than in men. The prognostic advantage of younger age and female gender remains after adjusting for the effects of all other well-established prognostic factors, which point to a more direct role of age and gender in melanoma. Further research is necessary in order to translate the survival advantage of younger persons and women into a survival benefit for all patients with melanoma.
- 6Changing trends in melanoma incidence and mortality [in French and English]. Health Rep. 1998; 10: 29–41 (English); 33–46 (French)., .
- 12Prognostic factors and natural history. BalchCM,HoughtonAN,SoberAJ,SoongAJ, eds. St. Louis, Mo: Quality Medical Publishing; 2003., , , .
- 15Should mass screening for skin cancer be targeted to middle age and older men? Melanoma Res. 2001; 11( supp 1): 150–151., , , et al.