Introduction

Malignant melanoma (MM) is an important global health challenge, and it is currently ranking number seven of the most common cancers worldwide. It is the fifth most common cancer among both sexes and the most aggressive skin cancer in Denmark [1, 2]. The incidence of cutaneous malignant melanoma is increasing worldwide, and the incidence has more than doubled during the past 25 years [3, 4]. There are approximately 2.738 new cases (1.333 men, 1.405 women) of malignant melanoma per year in Denmark [5]. Although the risk of melanoma increases with age, malignant melanoma is the most common cancer among young women and the second most common among young men (aged 15–39) [6]. MM carries the risk of both locoregional and distant metastases, although distant recurrences carry a poorer prognosis than local and regional metastasis [7]. The most common site of distant metastasis are lymph nodes, lungs, and the brain [8]. There are different factors such as tumor thickness, ulceration status, lymph node involvement and mitotic index which also affect prognosis and the risk of recurrences, although thickness is believed to be the most important factor [9, 10]. 10.6% of patients with MM will develop recurrences; therefore, detection of recurrences or new de novo MM in early stages is important for patient survival and prognosis [11]. The aim of follow-up schedules is to ensure early detection of recurrences as well as de novo MM.

International guidelines differ significantly internationally, but in Denmark patients are divided into three groups, low-risk (stage IA), intermediate-risk (stage IB-IIA), and high-risk (stage IIB-IIID), depending on their TNM and stage classification. The staging is based on UICC/AJCC TNM classification, 8th Edition. Denmark was the first country to implement routine imaging (FDG-PET/CT) as part of MM follow-up program for high-risk patients in 2016 [12]. The increasing incidence of MM will lead to a higher prevalence of eligible patients for follow-up programs. The increased demand will increase the burden on the Danish health care system which can result in delayed diagnosis and treatment of new patients due to long waiting lists. Therefore, it is important to evaluate the effectiveness of current guidelines and surveillances of MM.

This study aims to examine the output of Danish melanoma follow-up program with a focus on recurrences as well as de novo melanomas and the detection modalities used to identify them.

Material and methods

This study was a retrospective study based on prospectively collected patient data from the Danish Melanoma database (DMD) [13]. The study included all patients who were in the clinical follow-up program after primary treatment of cutaneous malignant melanoma at the Department of Plastic and Breast Surgery, Zealand University Hospital Roskilde from 01/01/2019 to 30/04/2023.

Exclusion criteria for this study consisted of:

  • Patients with disseminated melanoma (stage IV)

  • Patients who had recurrences before 01/01/2019

  • Patient with unknown primary tumor

  • Patients with unknown malignant or benign tumor

  • Patients with unmeasurable tumor thickness

A total number of 2456 patients were retrieved from the DMD, which of 408 were excluded, resulting in a final study population of 2048 patients. The included patients had between 1 to 14 registered follow-ups. Patients were followed until recurrences, end of follow-up period or death.

All the data retrieved from DMD were confirmed with manual review of the electronical medical records of the patients. Follow-ups after recurrence were not included in the study.

Demographical data on sex, age at the time of diagnosis, TNM classification, stages, metastasis, number of routine follow-ups and extra follow-ups, imagining modalities subcategorized in PET/CT or ultrasound was gathered at the time of diagnosis. The patients were categorized into stages based on their TNM-classification of coetaneous melanoma, AJCC 8th edition (2017) [14].

During each follow-up, the physicians registered whether the patients had any recurrences/de novo MM as well as whether it was detected by the patient’s self-examination, physicians, independent physicians, or routine imaging. Independent physicians were physicians who were not part of follow-up program, such as the GP; dermatologists, or from other departments of the hospital; whilst physicians were the physicians as a part of the follow-up program at the hospital. Cases of confirmed recurrences/de novo were registered just once, by the first detected modality. Patients were able to contact the department in case of any suspicious/new findings, which resulted in offers of extra follow-ups. All the extra/out of protocol follow-ups were also included in the study.

Patient follow-ups

The Danish clinical follow-up program for malignant melanoma was based on both TNM-classification and staging, according to the UICC/AJCC TNM classification, 8th Edition [14].

Based on the initial risk group placement, the patients were offered different follow-up programs.

  • Group 1: Patients with low risk which included stage IA patients, were offered one follows-up 3 months after the removal of the primary MM.

  • Group 2: Patients with intermediate risk which included stage IB and IIA patients, were offered follow-ups biannually for 5 years.

  • Group 3: Patients with high risk which included stage IIB, IIC, IIIA, IIIB, IIIC and IIID patients, they were offered follow-ups every 3 months for the two years, afterwards biannually for 3 years. In total 5 years of follow-up including a baseline PET/CT scan followed by routine PET/CT scans at 6-, 12-, 24- and 36-month follow-ups. Additionally, the patients were offered ultrasound scans of lymph node regions in case of positive sentinel nodes at all time points where PET/CT was not performed.

  • Group 4: Patients with ultra-high risk which included stage IV patients, they would initially be followed the same as patients in group 3, but there could be additional individual follow-ups. These patients were not included in the study.

After the end of 5-year follow- up program, the patients were recommended annual consultations with their dermatologists or general practitioner and regular self-examinations [15].

Data management

Data was retrieved from the Danish Melanoma Database (DMD) and stored on local servers hosted by Region Zealand. Data was supplemented by review of patient case files for additional information and quality assurance purposes.

Statistical analysis

RStudio (RStudio Team (2020). RStudio: Integrated Development for R. RStudio, PBC, Boston, MA URL http://www.rstudio.com/.) was used to perform a descriptive statistical analysis. A cox regression multivariate analysis was also conducted using RStudio. The Cox regression used to calculate hazard ratio (HR), while adjusted for sex, age, substages and time of recurrence. Stage IB is the reference, and stage IA and IIID are excluded from the analysis because of zero founded recurrences. P < 0.05 was considered significant for all tests. Tables and graphs were generated using Origin (Pro), Version Number (e.g., "Version 2022"). OriginLab Corporation, Northampton, MA, USA and Microsoft Excel version number (16.66.1).

Ethics

This study was conducted as a quality assurance project approved by the local Institutional Review Board.

Results

In total, 2048 patients and 7830 clinical follow-ups were included in the study. Among 2048 eligible patients, 53% were females and 47% were males. Median age was 64 years old (IQR: 59,0–79,2), the youngest patient was 15 years old and the oldest was 100 years old. The staging distribution revealed that 34% of patients were classified as stage IA, 37.8% as stage IB, 8.7% as stage IIA, 6.3% as stage IIB, 2.1% as stage IIC, 4.3% as stage IIIA, 3.1% as stage IIIB, 3.6% as stage IIIC, and no patients falling into stage IIID. Baseline characteristics are presented in Table 1.

Table 1 Baseline characteristics
Full size table

De novo

A total number of 38 de novo MM, 1.9% of total population were detected. De novo was mostly detected in stage IIIB patients (5.2%), Table 2. Among all registered de novo, 60.5% were detected by physicians during scheduled follow-ups, 28.9% by patients’ self-examination and 10.5% by independent physicians, Table 3.

Table 2 Confirmed recurrences and de novo mm in different substages
Full size table
Table 3 Confirmed recurrences and different detection modalities
Full size table

Recurrences

A total of 75 patients, constituting 3.7% of the study population, experienced recurrences. Staging distribution revealed varying recurrence rates: 0% in stage IA, 1.3% in stage IB, 6.1% in stage IIA, 14.7% in stage IIB, 18.6% in stage IIC, 2.2% in stage IIIA, 11.1% in stage IIIB, 24.3% in stage IIIC, and 0% in stage IIID, Table 2.

Among the identified recurrences, 58.4% were discerned through imaging modalities (PET/CT and ultrasound), followed by 16% detected by patients’ self-examination, 12.2% detected by independent physicians, 11.5% were detected by physicians during follow-ups, Table 3.

A total of 7830 follow-ups were recorded, with stage IB patients accounting for approximately 50% of the total follow-ups, Table 2.

Extra follow-ups, totaling 83, exhibited a 12% recurrence detection rate, Fig. 1.

Fig. 1
figure 1

The number of follow-ups (orange bar) at various time points and percentage of confirmed recurrences (blue bar) for each time point

Full size image

Stratifying by risk groups revealed an 8.5% recurrence chance in the intermediate-risk group (stage IB-IIA) and a 22% rate in the high-risk group (stage IIB-IIID) during extra follow-ups, Fig. 2.

Fig. 2
figure 2

Number of follow-ups divided in various time point (blue bar), and percentage of recurrences in each time point (yellow bar), in A) high-risk group (IIC-IIID) patients, B) intermediate-risk group (IB-IIA) patients. Time point expresses the time of performed follow-up while primary treatment is the baseline

Full size image

Figure 2 illustrates that 76% of recurrences occurred in high-risk group patients, contrasting with 24% in the intermediate-risk group. No registered recurrences were noted at 54-and 60-month follow-ups in high-risk patients, while intermediate-risk patients remained recurrence-free at 3rd, 18-, 24-, and 60-month follow-ups. The risk of recurrences in different time periods, categorized by risk groups, is depicted in Fig. 2.

The study incorporated 1208 scans, comprising 799 PET/CT and 409 ultrasounds. Notably, 40% of recurrences detected by PET/CT occurred before the 24-month follow-up, with none observed at 18-, 30-, and 60-month follow-ups. Ultrasound scans identified only three recurrences, all detected before the 18- month follow-up, with no occurrences thereafter, Fig. 3. This corresponded roughly to one recurrence found for every 136 performed ultrasound scan.

Fig. 3
figure 3

Number of performed scans (grey bar) A) PET/CT scans, B) ultrasound scans and percentage of recurrences (blue bar) in each time point. Time point expresses the time of performed follow-up while primary treatment is the baseline. It is to be noted that routine scans were performed at 6-, 12-, 24- and 36-month follow-ups, but as shown in Fig. 3, there are also registered scans in periods rather than mentioned ones. It is due to factor such as extra scans, delayed scans, etc. The scans have been registered to the nearest follow-up

Full size image

Figure 4 illustrates 65.3% of recurrences manifested before 24 months after the primary treatment, with a secondary peak at 36-month follow-up.

Fig. 4
figure 4

Number of recurrences by months from primary treatment

Full size image

A significant 78.7% of recurrences were identified before the 36-month follow-up.

Recurrence distribution showed 44% in regional lymph nodes, 37,3% distant metastases, and 19% locally. Routine PET-CT demonstrated superior efficacy in detecting lymph node and distant recurrences, while patient self-examination excelled in identifying local recurrences, Fig. 5.

Fig. 5
figure 5

Method of detection of recurrences, divided by sites of recurrences: distant, local, and regional lymph node. The percentages indicate the percentage of recurrences in various sites. Physician: The physician as a part of the follow-up program at the hospital. Independent physicians: other physicians such as GP; dermatologist or other departments at hospital; not part of follow-up program

Full size image

A multivariate analysis, calculating hazard ration while adjusting for sex, age, substages and time of recurrence revealed a significantly increased risk of recurrence in all stages except for stage IIIA patients, compared to stage IB (reference), Table 4.

Table 4 Multivariable Cox regression analysis of recurrence risk by substages
Full size table

Discussion

This study sought to comprehensively investigate the outcomes of the Danish melanoma follow-up program, with a specific emphasis on recurrences and de novo melanomas, as well as the modalities employed for their detection. Among the 2048 patients under scrutiny, 75 experienced recurrences, constituting 3.7% of the total study population. It is noteworthy that this recurrence rate appears lower in comparison to findings reported in other studies, where a broad range of 8.9% to 30% for recurrences has been observed [11, 16,17,18,19].

Importantly, contextual differences should be considered when interpreting these results. Our study uniquely included patients across all stages from IA to IIID, whereas many of the referenced studies primarily focused on either stage II or stage III patients. The divergence in reported recurrence rates may also be attributed to our exclusion criteria, which led to the removal of 408 patients, or may stem from variations in treatment and follow-up protocols across different countries. Thus, while our observed recurrence rate may appear comparatively lower, these variations underscore the importance of considering the diverse patient populations and methodologies employed across studies.

Different substages

Like other studies, recurrences differ in various substages. Our results suggested that recurrences are most common in stage IIIC patients (24.3%), followed by stage IIC (18,6%) and lastly stage IIB (14.7%). The results are in part consistent with Danish national observational cohort study, Helvind et al. 2023 [11]. They found sage IIID to be the most recurrent one, followed by stage IIIC and lastly stage IIC. This is due to the low number of stage IIID patients in our study which clearly could cause lower number of patients and possible recurrences.

It is though noticeable that stage IIIA patients had lower recurrence rate compared to stage IIB and IIC patients. The results are coherent with other studies [11, 20]. Winge-Main et al. 2023 found a poor cancer-specific survival in stage IIB and IIC patients compared to other stages. They mention that the poorer prognosis can be due to various factors such as melanoma type, tumor thickness, ulceration status, mitosis index, etc. Adjuvant therapy is another important factor that can influence recurrences in different stages. In the past few years adjuvant therapy has become part of treatment of stage III-IV patients, and this can probably explain the lower rate of recurrences in stage III patients compared to stage IIB-IIC patients. Ongoing trials are currently exploring the usage of adjuvant therapy in stage IIB-IIC patients, which can be beneficial for stage IIB-IIC patients (CT04099251, NCT04309409, and NCT03405155) and Keynote-716 (NCT03553836) [20].

Effectiveness of different modalities and sites of recurrences

Recurrences in this study were primarily detected in lymph nodes, followed by distant metastasis and local cutaneous recurrences, 44%, 37,3%, 19%, respectively. The results are consistent with Johannesen et al. [18]. The most common site of distant metastasis was lungs, followed by brain. The results tie with Lee et al. 2018, they also found lungs as the most common site of distant metastasis followed by the brain [19].

Among all recurrences, almost 55% were detected by PET/CT scans, 16% by patient’s self-examination, 12% by independent physician, 12% by physician and lastly, 4% by ultrasound scans. Even though physician was not the most effective method in finding recurrences, it was the most effective method in finding de novo MM. Approximately 61% of de novo MM were detected by physicians during follow-ups. These findings emphasize the importance of follow-up programs including PET/CT scans.

PET/CT was the most effective method in detecting recurrences in regional lymph nodes and distant metastasis. While the patient self-examination was the most effective one in finding local recurrences, which is in consistent with other studies [18, 21], this highlights the importance of self-examination which can have a big effect on detecting recurrences/ de novo in early stages by patients and the family [22, 23]. Danish health care also has a big focus on patient education and self-examination. Drabarek et al. 2023 also suggested that patient self-examination can be effective to reduce number of clinical follow-ups if the patients are assured that the self-examination in effective and safe [24].

While PET/CT showed to be an effective method in finding recurrences, ultrasound did not show encouraging results. Just 4% of recurrences were detected by ultrasound scan. It means ultrasound may be unnecessary for surveillance of melanoma patients. The results tie strongly with Gjorup et al. 2023 as they also found that routine ultrasound surveillance of node-positive patients may be unnecessary if regular cross-sectional imaging such as PET/CT is performed [25]. In our study, all three patients, that had ultrasound verified recurrence, would have had a PET/CT scan within a timeframe of three to nine months after the ultrasound verification.

Recurrences over time

According to our results, 65.3% of recurrences occurred within the first two years after primary treatment, which is consistent with prior studies [12, 21]. The second largest peak after 24-month occurred within 36-month follow-up, which is reasonable since a routine PET/CT or ultrasound would have been performed as a part of the follow-up program.

The guidelines within the Danish healthcare system are firmly rooted in evidence-based practices and consistently reflect the latest insights from both national and international research [26]. Notably, recent unpublished updates to the Danish melanoma guidelines have recommended a shift in follow-up frequency for patients in the intermediate-risk group (specifically, stage IB and IIA). Rather than the previous biannual schedule, the new guidelines advocate for annual follow-ups over a 5-year period. This adjustment is informed by a conscientious consideration of the notably low recurrence rates observed in stage IB patients, as corroborated by our study, where a mere 1.3% recurrence rate was identified in this particular subgroup.

Limitations

This study is subject to several limitations that warrant consideration. One notable limitation stem from our exclusion criteria, leading to the removal of nearly 400 patients initially classified as high-risk. Additionally, our 4-year follow-up time frame, while providing valuable insights, is relatively shorter compared to some larger cohort studies with follow-up periods extending up to 30 years [4, 17, 23, 27]. However, it is pertinent to note that existing research has indicated that a substantial majority of recurrences manifest within the initial two years post-primary treatment [28].

Furthermore, a noteworthy limitation lies in the relatively low number of patients in stage II-III, which may impact the generalizability of findings within these subgroups. Importantly, the study's duration is constrained to 4 years, and we acknowledge that not every patient was followed for the complete 5-year period. This temporal limitation resulted in an uneven distribution of follow-ups among participants, with some having only one or two follow-ups, while others contributed a more substantial proportion to the study dataset.

These limitations introduce the potential for missing patient follow-ups, which could, in turn, result in the underreporting of recurrences and contribute to a smaller effective study population. Finally, it is important to highlight that the study does not incorporate information regarding whether patients received adjuvant therapy, which represents a critical aspect that could impact recurrence patterns and outcomes. Addressing these limitations in future research endeavors will contribute to a more comprehensive understanding of melanoma recurrence dynamics.

Conclusions

In conclusion, the results showed the majority of recurrences were identified within the first three years of follow-up. The introduction of routine PET/CT scans have contributed to the timely identification of these events. The patients themselves find a high number of local cutaneous recurrences and this emphasized the importance of self-examination. Ultrasound was found to be a very low yield examination and its usefulness as a routine scan can be questioned. The overall findings underscore the importance of refining and optimizing melanoma follow-up programs according to staging to provide early detection and intervention. Future studies should also focus on whether this early detection and treatment leads to a survival benefit for the patients.