Abstract
BACKGROUND: The COVID-19 pandemic disrupted access to routine dermatologic care and may have delayed melanoma diagnosis and management. Evidence on the post-pandemic period and on hospital-based referral cohorts remains limited. We assessed melanoma presentations before, during and after the pandemic in three skin cancer centers in North Rhine-Westphalia, Germany. METHODS: We conducted a multicenter retrospective cohort study of inpatients with cutaneous melanoma grouped into Phase 1 (February 2017-February 2020), Phase 2 (March 2020-March 2023), and Phase 3 (April 2023-May 2024). The primary endpoint was Breslow tumor thickness (TT) among invasive melanomas, analyzed using multivariable log-linear regression adjusted for center, age, sex, anatomic site, and histologic subtype. Secondary endpoints included T category and AJCC stage distributions (including stage 0/Tis), macroscopic primary tumor specimen dimensions (area and volume; available cases), staging work-up and sentinel lymph node biopsy (SLNB) indicators, and exploratory laboratory parameters (LDH, S100, CRP) and dermal mitotic rate. RESULTS: We included 2960 patients (Phase 1: 1162; Phase 2: 1251; Phase 3: 547). Median TT among invasive melanomas was 1.1 mm (IQR 0.6-2.3), 1.1 mm (0.5-2.4), and 1.0 mm (0.5-2.3) across phases (p = 0.037). In adjusted models among invasive tumors, TT did not increase (Phase 2 vs. Phase 1: 0.97, 95% CI 0.90-1.04; Phase 3 vs. Phase 1: 0.94, 0.86-1.03). AJCC stage 0 decreased from 7.7% and 6.1% to 2.0%; adjusted OR Phase 3 vs. Phase 1: 0.24 (95% CI 0.13-0.46). Within invasive tumors, the distribution of T categories (T1a-T4) and AJCC stages I-IV was similar across periods. Among cases with available macroscopic primary tumor specimen dimensions, median area and volume were higher during and after the pandemic (area p = 0.030; volume p = 0.042), but period effects attenuated in models adjusted for TT. Exploratory analyses suggested a higher proportion of elevated LDH and a lower proportion of elevated S100 across periods, while CRP and dermal mitotic rate showed no clear period shift. CONCLUSIONS: In this large melanoma inpatient cohort, the pandemic period was not associated with thicker invasive melanomas after covariate adjustment. However, a persistent reduction in stage 0/Tis presentations in the post-pandemic period suggests ongoing disruption or shifting of early detection and referral pathways. Exploratory increases in macroscopic tumor dimensions may point to changes not captured by thickness alone, but require cautious interpretation given missingness and potential documentation effects.