Patterns in progression from early-stage melanoma to late-stage melanoma: implications for survivorship follow-up

Abstract

Aim: This study determined the characteristics of patients with early-stage melanoma (IA-IIA) who later had stage IV recurrence.

Patients & methods: We retrospectively examined 880 melanoma patients and identified those who progressed to stage IV disease from an initial early-stage (n = 50).

Results: We observed a median latent period of 4 years between early-stage diagnosis and metastatic disease. More patients (54%) developed metastatic disease 4 years or later from the initial diagnosis. 34% had regular dermatology appointments, and 30% had regular oncology follow-up. Lung and brain were the most common metastatic sites.

Conclusion: Long term monitoring beyond 4 years and a low threshold for performing symptom-guided imaging, particularly if pulmonary or neurologic symptoms occur, may be prudent after early-stage melanoma diagnosis.

Plain Language Summary

Early-stage melanoma has an encouraging five-year survival rate of 94%. However, a higher absolute number of fatal melanomas arise from thin rather than thick melanomas, due to the fact that thin, early-stage melanomas far outnumber melanomas that are initially diagnosed at thicker, more advanced stages. This study aimed to understand how and when patients with early-stage melanomas recur with metastatic disease. We examined the charts of 880 patients with melanoma who received care from oncology at a large academic medical center. We identified 50 patients who were initially diagnosed with early-stage melanoma (Stages IA-IIA) but then developed Stage IV melanoma. We measured time from first diagnosis to metastatic diagnosis, anatomic locations of metastases, if there were any symptoms at the time of metastatic diagnosis, and how patients had been followed after an initial early-stage diagnosis. We observed that half of patients developed metastatic disease more than 4 years after their early-stage diagnosis. In these patients, 34% had regular dermatology appointments, and 30% had regular oncology follow-up. Melanoma spread to the lung and brain most commonly. Shortness of breath and altered cognition were the most common symptoms that brought patients to medical attention at the time of stage IV diagnosis. This study demonstrates that the risk of melanoma recurrence exists even after 4 years for early-stage melanomas. Long-term symptom monitoring may be warranted beyond 4 years. Care providers should consider imaging when symptoms occur with no other explainable cause, especially neurologic changes or shortness of breath.

Plain language summary

Article highlights.

• Distant metastases from early-stage melanoma, while infrequent, are clinically meaningful and recurrence patterns have not yet been not fully described.

• This study derives from a panel of 880 patients with metastatic melanoma in order to describe the timing, presentation and clinical follow-up patterns at the time of recurrence or metastasis after an initial diagnosis of early-stage melanoma.

• Metastatic disease from early-stage melanoma occurred both early and late, with a majority of metastases observed to have occur after 4 years.

• Early-stage melanoma metastasized most frequently to lung and brain.

• Understanding these patterns may aid in surveillance planning and suggests a low threshold for brain and chest imaging in symptomatic patients, even with remote history of early-stage melanoma.

1. Introduction

Melanoma has a known capacity to metastasize even with a small primary tumor size, which raises many challenges in the approach to surveillance of early-stage melanoma [1]. Primary melanomas with an initial Breslow depth range of less than 1 mm to 1–2 mm account for nearly half of patients who die from melanoma, according to data from Queensland, Australia [2] and the USA [3]. For this reason, early detection is vital, especially in high-risk patients [4,5]. Many studies recommend routine follow-up for those diagnosed with an early-stage melanoma [6–8], even after definitive treatment. Notably these recommendations are driven by concern for detection of new primary melanomas and do not detail surveillance strategies specific to the detection of disease progression and distant metastases. The eighth edition of melanoma staging by the American Joint Committee on Cancer estimates a 96% 5-year survival rate for T1a and T1b melanomas, with survival rates of 98% for Stage IA, 94% for Stage IB and 88% for Stage IIA melanomas at 10-years [9]. However, some data suggest that the greatest mortality occurs in patients with melanoma that was early-stage at the time of diagnosis [10]. Additionally, the patterns and timeline for progression of early-stage melanomas to metastatic disease have not yet been well defined in the literature.

Single genetic point mutations in critical genes are known to cause early-stage melanoma. Late-stage and fatal melanomas are known to express increased aneuploidy and whole genome doubling in addition to point mutations in critical genes [11]. These genetic changes due to selective pressures are slow, consistent with the long latent period between early- and late-stage disease. Understanding the genetic pattern of recurrent melanomas may lead to methods of early detection, such as by circulating tumor DNA (ctDNA). There is no consensus for such long-term monitoring currently.

A latent period of two or more years has been observed between definitive treatment and recurrence in melanoma [12,13], with numerous factors known to contribute to this variation, including initial tumor depth and presence of ulceration [14]. By definition, early-stage melanomas do not manifest these high-risk features, thus routine survivorship surveillance is not uniformly done in this setting. Recommendations [6,14–16] for timing and duration of surveillance of early-stage melanomas are unclear. One group suggests regular follow-up for at least two years following early-stage melanoma diagnosis and treatment [14]. Another suggests that stage I patients follow-up every 6–12 months for the first five years followed by yearly follow-ups for life [6]. Meanwhile, the National Comprehensive Cancer Network's 2023 guidelines recommend a follow-up schedule that is based on the patient's personal and family health history [16]. Besides timing, an open question remains on whether such follow-up appointments should be under the care of a dermatologist, oncologist, or both. Multiple recommendations have been published [16,17]. The American Academy of Dermatology, for example, recommends that the follow-up and imaging schedule is variable based on the risk of disease recurrence [17]. These are important considerations given that this patient population does not have adjuvant medical treatment options aside from surgery, and they are excluded from clinical trials with adjuvant and neoadjuvant treatments.

This retrospective study seeks to describe the time interval between early-stage (IA-IIA) melanoma in cases which have progressed to stage IV disease and the associated clinical features of this type of metastatic progression. Within the medical oncology department of a single tertiary medical center, we examine time intervals, cancer surveillance, demographic and pathologic characteristics, genetic phenotypes, presenting symptoms leading to a diagnosis of stage IV disease, as well as sites of metastases. We analyze these data in the context of survivorship planning to assist patients and their cancer care providers to inform clinical care decisions for early-stage melanoma and enhance clinical discussions regarding the characteristics of metastatic disease within this largely understudied population.

2. Methods

This was a retrospective study of 880 consecutive melanoma patients from 2016–2020 under the care of a multi-physician practice at the Cleveland Clinic Taussig Cancer Center, a tertiary healthcare center. We identified new, non-de novo diagnoses of stage IV melanoma which progressed from an initial early-stage diagnosis. “Early-stage” was defined as stage IA-IIA under the 8th Edition Staging Manual by the American Joint Committee on Cancer (AJCC) [18]. Patients with mucosal and uveal melanoma were excluded.

Desired descriptive data of these patients included: demographics (age and gender), anatomic site of primary lesions, genetic tumor characteristics, dates of initial and progression diagnoses and follow-up history with both dermatology and oncology prior to progression. These data were then analyzed to calculate the mean and median age of early-stage melanoma diagnosis and time to stage IV progression. In addition, we also tabulated the most common organ systems with metastatic disease and the most common presenting symptoms at the time of a stage IV diagnosis. All documented patient stages were confirmed based on the AJCC 8th edition, which included patient histopathological (Breslow depth, ulceration, mitotic rate) and sentinel lymph node characteristics. Data were collected and secured in compliance with an IRB protocol (# 3164) approved and active with the Cleveland Clinic Institutional Review Board (IRB 3164). Verbal consent was not obtained because this was a retrospective chart review study for which data were collected under an IRB approved protocol.

3. Results

A total of 50 patients met the inclusion criterion of a stage IV melanoma with documented progression from an initial early-stage (stage IA, IB and IIA) diagnosis. Average age at time of initial diagnosis was 57.7 (median 60, range 21–68), and 62% (n = 31) were male (Table 1). An additional 8 patients were separately collected as “presumed early-stage”, due to appropriate but incomplete primary histopathologic data for an early-stage diagnosis. These patients were analyzed separately from the 50 confirmed early-stage patients that fully met inclusion criteria. For the presumed early-stage group, average age at time of initial diagnosis was 42 (median 38, range 31–59) and 50% (n = 4) were male (Table 1).

Table 1.

Characteristics of stage IV melanoma patients initially diagnosed as early-stage (IA-IIA) melanoma.

		Confirmed early-stage	Presumed early-stage
Total patients (n)		50	8
Stage at diagnosis	IA	22% (11)	75% (6)
	IB	46% (23)	25% (2)
	IIA	32% (16)	0% (0)
Mitotic index at diagnosis (per mm2)	Averagea	2.7 (43)	1.2 (5)
	Median	1	1
Age at diagnosis (yrs)	Averagea	57.7	42
	Median	60	38
Sex (M/F)		62%/38% (31/19)	50%/50% male (4/4)
Race (% non-Hispanic, white)		100%	100%
Regular dermatology follow-up		34% (17)	0% (0)
Regular oncology follow-up		30% (15)	25% (2)

^aMitotic index was unavailable for 7 confirmed early-stage patients and 3 presumed early-stage patients.

Of the 50 confirmed early-stage patients, 22% (n = 11) were stage IA, 46% (n = 23) were stage IB and 32% (n = 16) were stage IIA. For the presumed early-stage group, distribution by stage was 75% (n = 6) stage IA, 25% (n = 2) stage IB and 0% IIA. Average mitotic index was 2.7 (n = 43) for confirmed early-stage patients and 1.2 (n = 5) for presumed early-stage patients, which did not include 10 patients for whom mitotic index was not available at the time of initial diagnosis (7 confirmed and 3 presumed early-stage). Additional clinical and pathologic characteristics are listed in Table 1. There was not a relationship between the time to progression to stage IV disease and mitotic index (Supplementary Figure S1). The mean latent periods for progression from Stage IA, IB and IIA to stage IV melanoma were similar (Supplementary Figure S2).

In the confirmed early-stage patient cohort, median time from initial early-stage melanoma diagnosis to metastatic disease progression was 4 years (Table 3), with 46% (n = 23) diagnosed within 3 years, 22% (n = 11) between 4 and 6 years, 12% (n = 6) between 7 and 10 years, 8% (n = 4) between 10 and 12 years and 12% (n = 6) beyond the 12-year mark from their initial early-stage diagnosis (Figure 1). Immediately prior to stage IV diagnosis, a total of 30% (n = 15) and 34% (n = 17) maintained follow-up with oncology and dermatology, respectively (Table 1).

Table 3.

Time intervals between early-stage and stage IV melanoma diagnosis.

Time to stage IV diagnosis (yr)	Confirmed early-stage	Presumed early-stage
Average	6.1	12.3
Median	4	10.5
Range	1–24	5–25

Figure 1.

Distribution of recurrence intervals based on time from initial diagnosis of early-stage melanoma to stage IV progression.

In patients who met inclusion criteria, the two most common anatomic sites of early-stage primary lesions were the extremities (44%, n = 22) and back (21.6%, n = 11). The two most common sites of metastatic disease were lung (46%, n = 23) and brain (28%, n = 14) (Figure 2). Diagnosis of stage IV disease was prompted by systemic symptoms in 44% (n = 22) of patients, palpable masses or skin lesions in 42% (n = 21) of patients, pathologic skeletal fractures in 4% (n = 2) of patients and incidental imaging findings in 10% (n = 5) of patients. In confirmed early-stage patients, the most common systemic symptoms were shortness of breath (n = 9) and neurologic symptoms (n = 8), particularly altered mental status and vision changes (Figure 3).

Figure 2.

Distribution of metastatic locations in patients with early-stage melanoma after progression to stage IV.

Note: Many patients had multiple metastatic sites at the time of their stage IV diagnosis.

Figure 3.

Systemic symptoms in melanoma patients at the time of stage IV presentation.

Note: Many patients had more than one symptom at the time of stage IV diagnosis.

Incidental diagnosis of stage IV disease occurred due to imaging scans in 10% (n = 5) of patients (Table 2). Three of these scans were performed as a part of regular oncologic surveillance: one was a PET/CT to monitor non-hypermetabolic lung nodules from initial time of diagnosis, a second was a surveillance PET/CT ordered by an oncologist two years after diagnosis in an asymptomatic patient, and a third was a staging CT chest performed after resection of regional disease recurrence. Of the two other scans done outside of regular oncologic follow-up for melanoma, one was a CT of the chest done after an abnormal mammogram and the second was a CT angiogram performed as routine surveillance for an abdominal aortic aneurysm.

Table 2.

Presenting symptoms at the time of progression to stage IV melanoma for patients initially diagnosed with an early-stage (IA-IIA) melanoma.

	Confirmed early stage	Presumed early stage
Cause for diagnosis
▪ Systemic symptoms	44% (n = 22)	87.5% (n = 7)
▪ Palpable mass/skin lesion	42% (n = 21)	12.5% (n = 1)
▪ Pathologic fracture	4% (n = 2)	0%
▪ Imaging finding	10% (n = 5)	0%

Available somatic genetic profiling of common drivers in melanoma found 40% (n = 20) of melanomas in the study cohort were BRAF mutant. 20% (n = 10) had NRAS mutations, and 4% (n = 2) had KIT mutations (Supplementary Table S5). Genetic status of common driver mutations was unknown for 6% (n = 3) of these patients, and the remaining 30% were confirmed to be wild type for BRAF, NRAS and KIT. Castle Biosciences testing was unavailable for all patients in this study.

4. Discussion

To our knowledge, this study is the first to describe the clinical patterns of progression from early-stage to late-stage melanoma. Understanding these patterns can help clinicians better understand the natural history of stage IV melanoma arising from early-stage disease, in order to make appropriate surveillance decisions that consider common presenting symptoms for early-stage melanoma progression. These data also aid patients in understanding the natural history of early-stage melanoma and in appreciating the importance of surveillance in survivorship, with the goal of increasing education and ultimately reducing the anxiety and angst that can occur with an unexpected cancer progression.

Data presented here demonstrate that the time period between definitive treatment of an early-stage melanoma and stage IV recurrence can be short, with 46% recurring within three years (Figure 1). However, in the majority of patients who experienced metastatic recurrence from early-stage melanoma, recurrence happened after the first three years. This late recurrence pattern creates a logistic challenge for clinical surveillance. Furthermore, our data show that melanoma recurrence frequency was similar in each three-year interval after the first three years (Figure 1) and supports clinical observations that the risk of recurrence remains low but consistent over the first 10 years and never completely dissipates, as demonstrated by the 24-year latent period for one patient with a stage IV diagnosis observed in this cohort (Table 3). Very delayed recurrences, even for patients with favorable risk early-stage disease, have been described by others [19]. These data suggest that patients who have completed their initial post-treatment melanoma surveillance may derive benefit from regular assessments and education thereafter for the timely identification of systemic symptoms.

The results of this study support regular surveillance as an important element of melanoma survivorship. Notably only 33% of patients in our study had routine dermatology or oncology follow-up following their early-stage melanoma diagnosis (Table 1). At a minimum, patients diagnosed with an early-stage melanoma (IA, IB, or IIA) should have yearly full body skin exam with a dermatologist to assess for skin changes or lymphadenopathy indicative of recurrent disease. These suggestions may be of even greater pertinence to males in older age groups based on our, and others’, findings [20]. Per current National Comprehensive Cancer Network (NCCN) [16] and American Academy of Dermatology [21] guidelines, patients with stage IA or IB disease are weakly recommended (category 2A) to have follow-up physician visits every 6–12 months for five years, and then on an annual basis as clinically indicated thereafter. Our data emphasizes the utility of regular surveillance in those with early-stage melanoma. Furthermore, these data suggest that such monitoring may be most productive in the first three years, an observation window when metastatic recurrence appears to be relatively high for early-stage melanomas.

Impactful surveillance for systemic symptoms is more challenging to determine. In this cohort, systemic symptoms were common at stage IV diagnosis (Table 2). Lung and brain were the most common metastatic organ sites (Figure 2); correspondingly, shortness of breath and neurologic changes were the most commonly observed symptoms, one or the other occurring in 32% of stage IV diagnoses (Figure 3). Overall, 44% were symptomatic at presentation of stage IV recurrence. While other studies have described late recurrences from early-stage melanomas, this study specifically examines the anatomic locations and symptoms that were observed when stage IV melanoma follows an initial diagnosis of stage IA-IIA melanoma. Screening studies have traditionally not been economically feasible for all patients with a history of early-stage melanoma. However, these findings may be of more clinical relevance as new surveillance tools become available. It also provides a better characterization of when using these new technologies will provide the greatest benefit. Biomarkers that can predict higher risk of stage IV progression within early-stage cohort could narrow down a smaller subset that may benefit from regular surveillance with imaging [22–24]. While other authors have demonstrated that melanoma mitotic index is an important prognostic indicator [25], our data do not show a correlation between the time to progression to stage IV disease and the mitotic index of an early-stage melanoma (Supplementary Figure S1). These results in combination could be compatible in that mitotic index may be a better predictor for early recurrences, but not for late recurrences. Nor was there a difference between the mean latent period for progression to stage IV disease from stage IA, IB and IIA melanoma (Supplementary Figure S2). In the future, the development and refinement of less invasive prognostic and surveillance modalities may be available. Machine learning has been used to parse clinical and histopathologic features of an early-stage melanoma to estimate recurrence risk [26], and could be extended in larger datasets to distinguish separate factors which are characteristic of early and late recurrences. ctDNA may facilitate early detection strategies with less burden to a greater number of patients [27]. Another strategy might include detection of circulating markers of metastatic evolution in melanoma based on molecular changes such as copy number variation and whole genome doubling, both of which have been consistently observed in advanced melanomas [11].

Relatively large numbers of early-stage diagnoses and potential for decades-long length of risk for metastatic recurrence makes an emphasis on survivorship education an appealing strategy for early-stage melanoma care. Traditional survivorship education for patients with early-stage melanoma typically consists of education on primary prevention (i.e., minimizing unprotected sun exposure) and secondary prevention (i.e., routine dermatology follow-up). Based on the potential for late recurrences beyond typical surveillance windows, it may also be prudent to educate patients about the common symptoms of late-stage disease progression [28,29]. In our data, the most common signs and symptoms at the time of stage IV disease progression were respiratory and neurologic complaints (Figure 3). Physicians caring for these patients should have a low threshold to refer for chest or brain imaging if these symptoms arise in patients with a history of early-stage melanoma.

There are several important study limitations. First, this was a retrospective study. While our initial patient panel of 880 melanoma cases was large, many of these patients came to our tertiary cancer center after starting their melanoma cancer care and survivorship elsewhere, came to our center for only a one-time “second opinion,” or started care at our center and then continued their care or survivorship monitoring closer to home. In addition, some patients with early-stage melanoma lacked available information on sentinel lymph node biopsy, which was first introduced as an important staging tool in 1992 and gained traction thereafter [30]. In some cases, particularly the ones with longer latency, the standard practices of prior decades and sparsity of older records made it difficult to define the exact diagnosis date and staging, which is why we chose to analyze presumed early-stage melanomas as a separate cohort. Likely some early-stage patients have not been included due to lack of accurate diagnostic information and/or follow-up who have now progressed to stage IV disease outside of our healthcare system. Other patients were excluded because of incomplete early staging documentation. These data gaps are further exacerbated by extended latent periods between early- and late-stage diagnosis. Despite these gaps, a high percentage of stage IV diagnoses originated as early-stage melanoma. From 218 total cases of stage IV melanoma, 58 cases were initially diagnosed as early-stage and presumed early-stage melanoma (26.6%), where 160 cases (73.3%) were initially stage IIB and greater disease. In contrast to the Whiteman et. al. cohort [31], where progression to stage IV from early-stage cases were numerically higher than from late stages, our early stage-to-stage IV cohort was numerically smaller than from stage IIB or greater. To our knowledge, all the patients in this analysis had only one primary melanoma. We cannot account for how many patients might have had a second primary melanoma that regressed and was not evident in the clinical history. Any such cases would lead to an overestimate of years between early-stage disease and metastatic disease.

In summary, these data demonstrate that metastatic disease recurrence after an early-stage diagnosis often occurs within the first three years, affects the brain and lungs most frequently and in over half, only occurs after a long (>4 years) latency period. Based on our data, we would recommend that early-stage patients be educated on the risks and symptoms of metastatic recurrence and encouraged to maintain regular physician follow-up at least consistent with expert guidelines. Patients should also have education about common symptoms of recurrence, such that they are better equipped to recognize alarming symptoms and present for prompt examination if changes arise. Physicians, likewise, should recognize that recurrent melanoma may have a very long latency period after early-stage melanoma and have a low threshold for imaging, particularly if concerning neurologic or respiratory symptoms are identified. Most importantly, proper patient education after an early-stage melanoma diagnosis can lead to a shared decision-making discussion on melanoma surveillance that is well informed, fits the patient's risk tolerance profile and may ultimately lead to earlier diagnosis and improved outcomes of stage IV melanoma. It is our hope that these findings will aid in the optimization and institutionalization of early-stage melanoma survivorship planning to better serve this understudied patient population.

5. Conclusion

Long term clinical monitoring for systemic relapse beyond 4 years may be considered for patients with early-stage melanoma. Providers should have a low threshold for performing imaging studies when neurologic or respiratory symptoms develop in these patients.

Supplemental material

Supplemental data for this article can be accessed at https://doi.org/10.1080/20450885.2024.2424708

Author contributions

All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by DJS, BS, RP and PF. DJS prepared the tables and figures. The first draft of the manuscript was written by DJS and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Ethical conduct of research

Data were collected and secured in compliance with an IRB protocol (# 3164) approved and active with the Cleveland Clinic Institutional Review Board (IRB 3164). Verbal consent was not obtained because this was a retrospective chart review study for which data were collected under an IRB approved protocol.