Mustafa Aktaş, Senem Çengel Kurnaz, Emre Demirel, Ozan Çolak, Emel Tahir, Esra Kavaz Uştu, Mehmet Çelebi
Surgical margin status, perineural invasion and nodal stage were consistent, independent determinants of prognosis that identify high-risk patients; clear margins and accurate nodal assessment should be prioritised, though the retrospective design precludes causal inference.
PURPOSE: To identify the independent prognostic determinants of overall, disease-specific and recurrence-free survival in cutaneous squamous cell carcinoma (cSCC) of the head and neck in a single-centre cohort.
METHODS: Eighty patients who underwent surgery for head and neck cSCC between January 2008 and January 2023 were reviewed retrospectively. The effect of clinical, surgical and pathological variables on three endpoints was assessed with univariate Cox regression. To avoid overfitting, the multivariable model for each endpoint was pre-specified with a small number of covariates scaled to the number of events. Kaplan-Meier curves were compared with the log-rank test.
RESULTS: Mean age was 71.3 years; 85% were male. Mean follow-up was 46.3 months; 47 patients died (23 from disease) and 41.2% developed recurrence. In pre-specified multivariable models, surgical margin, perineural invasion and nodal stage were independent determinants of all three endpoints. For disease-specific death, the adjusted hazard ratios (aHR) were 4.39 (95% CI 1.77-10.86) per margin category, 5.05 (95% CI 1.99-12.79) for perineural invasion, and 2.43 (95% CI 1.48-4.00) per nodal-stage level. Older age was independent for overall survival only (aHR 1.06; 95% CI 1.03-1.09). The proportional-hazards assumption held, and on internal validation the optimism-corrected c-index remained high (0.74-0.85). Competing-risks analyses confirmed surgical margin, perineural invasion and nodal stage for disease-specific death, while perineural invasion was the most robust recurrence predictor.
CONCLUSION: Surgical margin status, perineural invasion and nodal stage were consistent, independent determinants of prognosis that identify high-risk patients; clear margins and accurate nodal assessment should be prioritised, though the retrospective design precludes causal inference.