FormalPara Key Summary Points

Why carry out this study?

It remains unclear whether patients with a history of multiple primary melanomas (MPM) have a worse survival prognosis compared with single primary melanoma (SPM) patients. This study investigates the demographics, histological features, and survival of MPM versus SPM patients to determine if (1) some subgroups are more at risk to develop MPM, (2) MPM patients present a worse survival compared with SPM patients, and (3) adaptations of the follow-up protocols of CM patients are required.

What was learned from the study?

A younger age at first diagnosis was identified as a risk factor for developing MPMs. Furthermore, the results sustain the hypothesis that MPM patients might have a worse overall survival compared with SPM patients, and that an increased follow-up is recommended.

Introduction

Cutaneous melanoma (CM) represents a significant health burden, especially in fair-skinned populations [1]. The incidence of CM is steadily increasing throughout the world [2]. In Belgium, CM represents the fourth and sixth most frequent cancer in women and men, all ages considered, respectively [3]. Approximately 3500 new CM cases are diagnosed each year, but these numbers are steadily rising, with an annual increase of approximately 5% [4].

Patients with a history of CM are at risk of developing subsequent melanomas. The risk of developing a second primary melanoma in CM patients is higher than the risk of developing a first CM in the general population [5]. The frequency of multiple primary melanomas (MPM) ranges from 0.2% to 12.7% of CM patients [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24].

Although different scientific societies have published guidelines (including European Society Medical Oncology (ESMO), European Organization for Research and Treatment of Cancer (EORTC), European association of dermato-oncology (EADO), European Dermatology Forum (EDF), and American Academy of Dermatology (AAD)), there is no unanimous consensus for the dermatological follow-up of CM patients, which is mostly based on the initial tumor staging and not on the risk of subsequent melanoma [25,26,27]. Standard follow-up in our center is, regardless of staging, every 3 months for the first 2 years, followed by twice a year for the next 3 years, and then annually for the long term.

To investigate if MPM patients require a specific follow-up regimen, this study evaluated whether SPM versus MPM patients present different survival rates, and compared the demographics and histological features of SPM and MPM patients.

Material and Methods

Patients

The local university ethics committee approved the design of the study (ref. 2017/334). From the melanoma database from the Dermatopathology Department of the University Hospital Centre of Liège, all patients diagnosed with CM between 1 January 2000 and 31 July 2019 were selected. Among those, MPM patients were identified by matching the names appearing twice or more in the histological records of melanoma. Both in situ and invasive melanomas were included in this study. Metastatic and recurrent CM were excluded.

Demographics

From all the MPM patients, date of birth, gender, age at the diagnosis of the first and subsequent melanomas, the respective anatomical locations, and the delay between the first and subsequent melanomas were recorded. The vital status and eventual date of death of all the SPM and the MPM patients at the end of the study (31 July 2019) was identified using the civil registries.

Histology

All the histological slides of the primary and subsequent primary melanomas were reexamined by one dermatopathologist according to the latest staging guidelines, including melanoma subtype, Clark level, Breslow’s thickness, mitotic activity, tumor-infiltration lymphocytes (TIL), presence of microscopic satellites, lympho-vascular embolization, neurotropism, presence of ulceration, melanomas developing on previous existing melanocytic lesions, and signs of regression [28]. Furthermore, the expression of immunohistochemical markers, including HMB45, P16, SOX10, Ki-67, and Melan-A were reassessed in the primary and subsequent melanomas.

As the data of the dermatopathology files of the SPM were not always available, we used the mean Breslow index of SPM of 0.90 (IQR 0.52–1.80) mm as the reference population [24].

Survival

Survival probabilities of SPM and MPM patients were represented using Kaplan–Meier curves, yielding global 5-year survival rates. Subsequently, univariate and multivariate Cox regressions were performed, with and without a time-varying covariate to assess the impact of the survival bias of the immortal time for the MPM patients, yielding a hazard ratio (HR) for death risk [24].

Statistics

Continuous variables are presented as means, standard deviations (SDs), or medians (quartiles). Qualitative variables are presented as frequency tables. Quantitative parameters were compared using Student’s t-tests (Kruskal–Wallis test) and qualitative parameters were compared using chi-squared tests.

Results

Demographics

The total number of patients and MPM patients, gender, vital status at the end of the study, age at first melanoma, time interval between first and second melanoma, and duration of follow-up since first melanoma diagnosis are presented in Tables 1 and 2. Out of the cohort of 95 MPM patients (females: 66.3%; males: 33.7%), 81 presented with 2 (85.2%), 10 with 3 (10.5%), 3 with 5 (3.2%), and 1 with 8 primary melanomas (1.1%).

Table 1 Total number of patients and MPM patients, gender, and vital status at the end of the study
Full size table
Table 2 Age at first melanoma, duration of follow-up since first melanoma diagnosis, and time interval between first and second melanomas
Full size table

Among the second melanomas, 65 (68.4%) were diagnosed during the first 5 years following diagnosis of the first melanoma, and 30 (32.6%) were diagnosed after more than 5 years. Figure 1 shows the number of cases per year diagnosed after the first diagnosis. Moreover, four patients had synchronous tumors, defined as less than 30 days between excision of first and second lesions. Regarding tumor localization on the body, the most common location was the lower extremities, followed by the trunk, then other sites.

Fig. 1
figure 1

Number of cases per year diagnosed after the first diagnosis

Full size image

Histology

Among the 95 MPM patients, 29 (30.5%) presented an in situ melanoma as their first melanoma. Among the second melanomas, 38 (40%) were in situ melanomas. Out of the 29 first in situ melanomas, 14 (48%) developed a second invasive melanoma. Moreover, there were significantly more in situ second melanomas in the first 5 years (n = 31) than > 5 years (n = 7).

The mean Breslow thickness of the first melanomas (0.84 mm; min: 0; max 16 mm; SD 1.77) was greater than that of the second (0.38 mm) and third (0.33) invasive melanomas.

The mean Breslow thickness of the second melanomas diagnosed in the first 5 years after the first one was 0.30 mm, compared with 0.55 mm after the first 5 years (Fig. 2). Concerning the other histological features, including melanoma subtype, Clark level, mitotic activity, tumor-infiltration lymphocytes (TIL), presence of microscopic satellites, lympho-vascular embolization, neurotropism, presence of ulceration, melanomas developing on previous existing melanocytic lesions, and signs of regression, no differences were found between the first and subsequent melanomas.

Fig. 2
figure 2

Mean Breslow thickness of the second melanomas diagnosed during or after the first 5 years

Full size image

MPM patients with a greater Breslow thickness for the first melanoma had a significantly decreased mean interval between first and second tumor (Fig. 3).

Fig. 3
figure 3

Mean time interval between first and second melanoma according to the Breslow thickness of the first melanoma

Full size image

Histologically, most first tumors were superficial spreading melanomas (64.2%), followed by in situ melanomas (28.4%), nodular melanomas (6.3%), and lentigo maligna melanomas (1.1%). The distribution for the second melanomas was similar, except for lentigo maligna melanomas being as frequent as nodular melanomas.

In MPM patients, the percentages of the first and second melanoma developing on preexisting naevi were 13% and 12%, respectively.

Survival

The overall 5-year survival rates of the total cohort of MM patients (n = 3853) was 91.5%. The overall survival rate of the SPM group (n = 3758) and the MPM group (n = 95) were 91.4% and 96.8%, respectively (Fig. 4).

Fig. 4
figure 4

Global 5-year survival rates of SPM and MPM patients

Full size image

The univariate and multivariate (adjusted for sex and age) Cox regressions revealed no statistically significant differences between death risk for SPM and MPM patients, respectively (p = 0.26, p = 0.99).

Taking account of the phenomenon of immortal time, the univariate Cox regression did not reveal a statistical difference (HR 1.219, p = 0.52), whereas the multivariate Cox regression, adjusted for age and sex, revealed a tendency towards a worse prognosis with a HR of 1.683 (p = 0.09).

Discussion

The incidence of MPM is in our population was 2.5%, similar to the 0.2–12.7% reported in other series [5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24].

Contradictory results regarding the sex ratio in MPM patients were described, but our population presented a significant preponderance of female patients [5,6,7, 9, 11,12,13,14,15, 17, 19,20,21,22, 24, 29]. In fact, the sex ratio seems to correlate with that of the SPM population, as published in previous studies [6, 11, 14, 15, 22]. Moreover, 10 out of the 14 (71%) patients with three or more melanomas were females, contrasting with previous observations stating that male patients tended to have a greater number of primary melanomas [7, 13].

Regarding the age at first diagnosis of MPM patients, our results indicated a younger age (male = 54.6 years, female = 44.8 years) compared with the age at first diagnosis in SPM patients (male = 63.3 years, female = 58.2 years) [6, 12]. However, other authors described the opposite trend [13, 19, 21, 22, 24, 29]. This may be explained by the higher number of patients included with a stronger statistical significance.

In accordance with previous reports, the majority of MPM patients developed two primary tumors [6,7,8,9,10,11,12,13,14, 17, 19, 20, 22,23,24].

Again, as mentioned in previous studies, the subsequent melanomas were more likely to be in situ or thinner than the first one. [6,7,8,9,10,11,12, 14,15,16,17,18,19, 22,23,24] This phenomenon has been linked to closer follow-up visits and self-surveillance [11, 12, 16, 22]. In addition, one study failed to prove that MPM patients have a less aggressive tumor biology [22]. Furthermore, 22 patients (23%) had a second melanoma with a higher Breslow thickness than the first melanoma, consistent with previous results [6, 14, 15]. This may be related to the fact that some patients evade medical surveillance by noncompliance or relocation.

Of the second subsequent melanomas, 28.4% of the total number of MPM detected were surgically excised during the first year, rising to 38.9% in the first 2 years, and to 68.4% within 5 years. The remaining 31.6% were excised after 5 years. This is in line with previous studies reporting that the majority of second subsequent tumors were detected within the first 2 years. [6, 8, 14, 15, 17, 18, 20] Only four patients (4.2%) had synchronous lesions, defined as multiple melanomas excised within 3 months, contrasting with previous findings, ranging from 13% to 45%. [6,7,8, 10,11,12,13,14,15, 17, 20, 23]

Regarding the anatomical region, the lower extremities were the most common site in women for SPM and MPM, while the posterior trunk was the most frequent in men. Only one article revealed the head and neck area as the most commonly involved site [19]. The anatomical site of the first and second subsequent melanomas was the same in 44% patients, similar to previous results. [7, 10, 11, 14, 15, 22]

Our study revealed that the Breslow’s thickness of second subsequent melanomas diagnosed during the first 5 years was lower (0.3 mm) than those diagnosed after 5 years (0.55 mm). Furthermore, there were more in situ melanomas identified during the closer follow-up schedule (< 5 years) (31/65; 47.7%) than afterwards (7/30; 23.3%). Again, these demonstrate the benefit of a closer follow-up system.

Another interesting finding was that the thicker the primary melanoma was, the sooner the subsequent was identified (Fig. 3). This could be linked to a better compliance to the follow-up by the patient, as an initial prognosis of a thicker lesion probably renders the patient and the dermatologist more observant.

Concerning the overall survival of MPM patients, two major hypotheses are given: first, survival could be better for SPM patients as follow-up is usually closer and there might be an immunological surveillance, as has been described for patients with three or more primary melanomas [13]. On the contrary, survival could be worse as there could be an increased risk for metastasis. [29] Our results evidenced only a slight difference in overall survival (91.4% versus 96.8%) at 5 years. Avoiding the methodological bias of the “survival bias” or “immortal time,” the statistical analysis also showed a tendency of worse survival for MPM patients (HR 1.683, p = 0.09), in accordance with previous publications [24, 30]. Another study used delayed-entry methods and also found a worse survival for MPM patients [29]. Finally, it has also been suggested that including in situ melanomas in the survival studies does not influence the survival of MPM patients, and that it might dilute the opportunity to detect a real effect [31].

One limitation of these results could be that some subsequent melanomas could have been excised in other centers, underestimating the number of MPMs.

Another limitation is that data relating to familiarity, the presence of predisposing mutations, BRAF mutational status, and general endogenous and exogenous risk factors in both groups of patients were not available in our study cohort.

A final limitation is that only the overall survival was studied and not the melanoma-related death rate. Having MPMs could decrease the overall 5-year survival not only due to melanoma, but also to other medical origins or a genetic predisposition for other cancer types; data that were not available in our database.

Conclusions

In our cohort, a younger age at first diagnosis was identified as a risk factor for developing subsequent melanomas. Furthermore, our results sustain the hypothesis that MPM patients might have a worse overall survival compared with SPM patients, and that an increased follow-up is recommended.

Guidelines should take this phenomenon into consideration, and patients with cutaneous melanoma regardless of the tumor thickness should be followed-up for life. Precise follow-up regimens are still to be determined. In the future, genetics may be added as an additional tool to identify patients at risk of MPM.