Cutaneous malignant melanoma (MM) constitutes a relatively small fraction of all skin cancers, yet its global incidence has been steadily increasing. In 2020, approximately 325,000 new cases and 57,000 deaths were reported worldwide [1]. In Korea, the incidence of melanoma has also risen among cutaneous malignancies in recent decades [2]. While excessive ultraviolet (UV) exposure is recognized as a major risk factor for melanoma, Asian populations show a predominance of acral sites—the soles, palms, and nail units—where the relationship to UV exposure is minimal [3]. Multiple epidemiological studies have implicated age, sex, anatomic site, comorbidities, and lifestyle factors in the development and clinical characteristics of melanoma; however, data from Asian populations remain limited [4-6]. The present study investigates the clinical implications of temporal changes in a 22-year cohort of 960 patients with skin cancer treated at a single tertiary care center, with particular emphasis on those diagnosed with MM.

Over the period 2000–2022, a total of 960 cutaneous malignancies were surgically treated. Group composition was as follows: BCC (455 cases, 47.4%), SCC (425 cases, 44.3%), MM (56 cases, 5.8%), and others (e.g., sarcoma; 24 cases, 2.5%) (Table 1).

Anatomic site distributions differed significantly among tumor groups (chi-square p<0.001). BCC and SCC were predominantly located on the head and neck (BCC: 427/455, 93.8%; SCC: 320/425, 75.3%). In contrast, MM showed a more even distribution: head and neck, 8/56 (14.3%); trunk, 17/56 (30.4%); non-acral extremity, 12/56 (21.4%); and acral, 19/56 (33.9%) (Table 1). Sex distribution also varied significantly across groups (p=0.005), with a higher male proportion in SCC (68.0%) compared with BCC (59.6%), while MM exhibited near parity (50.0%). Mean ages were 77.66±12.30 years for MM, 81.67±12.50 years for BCC, 86.17±11.46 years for SCC, and 68.75±23.67 years for other malignancies (ANOVA p<0.001; Kruskal-Wallis p<0.001).

Among the 56 MM cases, 25 occurred during period A (2000–2017) and 31 during period B (2018–2022) (Table 2). The anatomic site distribution changed significantly over time (chi-square p=0.043). In Group A, the proportions were 28.0% for head/neck, 20.0% for trunk, 24.0% for non-acral extremity, and 28.0% for acral sites; in Group B, the corresponding proportions were 3.2%, 38.7%, 19.4%, and 38.7%, respectively. These results indicate a decline in head/neck involvement and increases in trunk and acral lesions (Table 2, Fig. 1).

Sex and age did not differ significantly between the two periods (each p>0.05). The proportion of male patients increased from 40.0% in Group A to 58.1% in Group B (Fisher p=0.282). Mean ages were 66.2±11.4 years in Group A and 68.9±11.7 years in Group B (Mann-Whitney p=0.531; Welch’s t-test p=0.397; mean difference B–A=+2.65 years, 95% CI −3.58 to +8.87) (Table 2).

Analysis of annual incidence rates revealed 1.39 cases per year (95% CI, 0.90–2.05) for 2000–2017 and 6.20 cases per year (95% CI, 4.21–8.80) for 2018–2022. Considering all 56 cases, the exact binomial test showed a significant difference between periods (p<0.001), with a RR of 4.46 (95% CI, 2.64–7.56), indicating an approximately 4.5-fold higher annual incidence in the recent 5-year period compared with the preceding 18 years (Fig. 2). By contrast, Group A exhibited a 3-year recurrence rate of 36.0% and a 3-year survival rate of 76.0%, whereas Group B showed a 3-year recurrence rate of 35.5% and a 3-year survival rate of 80.6%. Differences between periods were not statistically significant (Fisher exact test: p=1.000 for recurrence; p=0.750 for survival).

Comparisons of comorbidities and social history between Groups A and B showed that hypertension was more frequent in B (32.0% vs. 64.5%: OR [B vs. A], 3.86; 95% CI, 1.26–11.80; p=0.031), and diabetes also increased (8.0% vs. 35.5%: OR, 6.33; 95% CI, 1.25–32.01; p=0.024). Pulmonary disease appeared only in Group A (4 cases) and not in B (p=0.034), but after zero-cell correction, the 95% CI included 1 (corrected OR, 0.08; 95% CI, 0.00–1.48), warranting cautious interpretation. Cerebrovascular disease, prior cancer, bone disease, and heart disease did not differ between the two periods (each p>0.05). Smoking and alcohol use tended to be higher in Group B, although these differences were not statistically significant (each p>0.05) (Table 3).

In this 22-year, single-center cohort, BCC (47.4%) and SCC (44.3%) constituted the majority of cutaneous malignancies, while MM accounted for 5.8%. These proportions align with global and domestic epidemiological patterns in which BCC and SCC dominate overall incidence [7,8]. Consistent with patterns of chronic UV exposure, BCC and SCC were concentrated in the head and neck region (BCC, 93.8%; SCC, 75.3%) [7,9]. Additionally, although uncommon, SCC may arise in non-UVrelated contexts, including post-acupuncture keloid scars or following kidney transplantation [10,11]. In contrast, MM demonstrated a more heterogeneous distribution (head/neck, 14.3%; trunk, 30.4%; non-acral extremity, 21.4%; acral, 33.9%), reinforcing that melanoma pathogenesis cannot be attributed solely to UV exposure [9,12-14].

Sex and age also differed significantly across tumor groups (chi‑square p=0.005; ANOVA/Kruskal-Wallis p<0.001). SCC had the highest male proportion (68.0%), BCC had 59.6% males, and MM was balanced (50.0%). Mean age was highest in SCC, followed by BCC, and lowest in MM, suggesting relatively earlier diagnosis for MM compared with keratinocyte cancers. These patterns parallel the head/neck predilection and male predominance of UV‑related cancers (potentially reflecting occupational/outdoor exposure) and the more balanced sex distribution and broader anatomic spectrum of MM with its non‑UV pathways. Clinically, combining age, sex, and site may inform surveillance and biopsy thresholds in plastic surgery practice (e.g., low biopsy thresholds for head/neck keratinocyte lesions in older men, and heightened vigilance for acral lesions across ages and sexes).

MM arises from malignant transformation of melanocytes through both UV-related and non-UV mechanisms (e.g., acral and mucosal pathways). The manifestation varies by anatomic site, but most lesions appear as a brown-black mixture; rarely, they present as a plaque-type blue nevus [15]. Histopathological subtypes include superficial spreading, nodular, lentigo maligna, and acral lentiginous melanoma, each with distinct molecular features [9,12-14]. This biological heterogeneity influences anatomic distribution, diagnostic timing, and treatment approaches, all of which are clinically relevant to reconstructive planning. In Asian populations, acral melanoma is comparatively prevalent [9,12], and our cohort likewise demonstrated a high proportion of acral cases (33.9%). Acral lesions often present with atypical pigmentation, ulceration, or nail changes, which may lead to delayed diagnosis, greater Breslow thickness, ulceration, and poorer prognosis. Early recognition in primary and secondary care and a low threshold for biopsy of acral lesions are therefore essential.

In this study, all surgical procedures were performed by a single surgeon over the 22-year period. Preoperative positron emission tomography-computed tomography (PET-CT) was routinely obtained to evaluate metastatic disease, and wide excision with a uniform 2-cm safety margin was performed regardless of tumor stage. When regional lymph node metastasis was identified, lymphadenectomy was performed; sentinel lymph node biopsy was not conducted as a separate procedure. Patients were followed with PET-CT at 6-month intervals for 3 years. Chemotherapy and/or radiotherapy were administered in cases with confirmed metastasis.

We defined study periods by comparing the most recent 5 years with the preceding interval to capture contemporary melanoma trends, while targeting similar sample sizes per group to enhance statistical power. The robustness of the findings was assessed using alternative temporal cut points in sensitivity analyses. In the period-stratified analysis of melanoma (Groups A and B), neither sex nor age differed significantly (each p>0.05). The only significant change in site distribution was a reduction in head and neck lesions from 28.0% to 3.2% between periods A and B, accompanied by increases in trunk and acral lesions to 38.7% each (chi-square, p=0.043). Regarding comorbidities, hypertension (64.5%) and diabetes (35.5%) were significantly more frequent in period B (p=0.031 and p=0.024, respectively), and this period effect persisted after adjusting for age, sex, and site, suggesting a higher metabolic disease burden in the recent case mix. Possible contributing factors include shifts in referral or transfer patterns, increased awareness among patients and clinicians leading to more acral referrals, and changes in healthcare utilization. As health awareness has increased and screening intervals have shortened, the diagnosis and initiation of treatment for baseline comorbidities such as hypertension and diabetes have generally occurred earlier. Similarly, head and neck lesions that were previously dismissed as benign nevi or dermatitis now appear more likely to be biopsied or excised at premalignant stages as patients present earlier. In contrast, lesions on the trunk and acral sites—areas less easily self-examined—are more likely to be detected at more advanced stages, which may partly explain the observed redistribution of anatomic sites. Together, these patterns suggest an interplay between system-level detection or referral sensitivity and true epidemiologic change. Additional system-level confounding may also be present. Heightened health awareness has increased care-seeking even for minor conditions, and during the COVID-19 pandemic, along with widespread adoption of electronic medical record systems, referrals and transfers from primary and secondary facilities to tertiary centers became more frequent as additional symptoms were identified after initial visits for minor illnesses. Whether this represents a true epidemiological shift or a pattern influenced by systemic and behavioral factors warrants follow-up multicenter, longitudinal analyses. Smoking and alcohol use also trended upward, although differences were not statistically significant (p=0.737 and p=0.116, respectively). Given the small cell counts and multiple comparisons (nine tests), chance variation cannot be excluded. Nonetheless, because metabolic and behavioral factors often cluster with socioeconomic determinants, these findings should be interpreted in the context of changing case mix, pandemic-related lifestyle shifts, and heightened sensitivity of clinical history taking and coding practices.

The annual incidence rate of melanoma in period B (6.20 cases per year) was approximately 4.5 times that in period A (1.39 cases per year: RR, 4.46; p<0.001). National data likewise indicate a long-term rise in skin cancer and melanoma incidence in Korea, consistent with our directional findings. However, single-center increases may also reflect aging demographics, greater access to biopsy and screening, enhanced electronic medical record documentation, and expanded referral inflows. Future investigations should incorporate population standardization, adjustment for referral volume, and year-by-year analyses to distinguish the true magnitude of incidence growth from detection and inflow effects.

Strengths of this study include its large, department-based cohort spanning 22 years, enabling direct comparison of keratinocyte cancers and melanoma within a single plastic surgery practice. The study also benefits from explicit rate calculations adjusted for exposure time and use of exact tests demonstrating a recent rise in melanoma caseload. The high proportion of acral and nail unit disease has direct implications for both early recognition and reconstructive planning, including graft versus flap selection, onychoplastic strategies, and determination of amputation levels. Limitations include its retrospective, singlecenter design, modest melanoma sample size (n=56), absence of prognostic variables, and limited control for referral flows, regional demographics, and pandemic-related effects. In addition, key pathological parameters, such as Breslow thickness and AJCC (American Joint Committee on Cancer) stage, were missing or incomplete for a subset of earlier cases in the institutional archive, precluding reliable statistical analysis. Consequently, further studies incorporating complete pathology and staging data are warranted. Prospective, multicenter studies incorporating population standardization and outcome and molecular data are warranted.

In this single-center cohort, melanoma accounted for 5.8% (56/960) of cutaneous malignancies, with a notably high acral proportion. Over time, the anatomic distribution shifted from head and neck toward trunk and acral sites, and the annual melanoma caseload in the most recent 5 years was significantly higher than in the preceding 18 years. Hypertension and diabetes were more common in recent years, accompanied by nonsignificant upward trends in smoking and alcohol use. These findings underscore the need for earlier detection and diagnosis of acral and nail unit lesions and for standardized reconstructive pathways in plastic surgery. Further multicenter, prospective research with population adjustment and inclusion of prognostic endpoints is needed to validate these observations.