Abstract
BACKGROUND: The last revision of the American Joint Committee on Cancer (AJCC) Cancer Staging Manual included a specific system for cutaneous squamous cell carcinoma (CSCC) of the head and neck. Here, we assessed the prognostic performance of six candidate modified T-classification models in head and neck CSCC patients. METHODS: Analysis of 916 patients with head and neck CSCC given treatment with curative intent at The University of Texas MD Anderson Cancer Center between 1995 and 2019 was performed. The main outcome was disease-specific survival (DSS), and the impact of depth of invasion (DOI) was analyzed using multivariable regression models. Candidate models were developed using the optimal DOI cut points for each AJCC T classification based on goodness of fit of the model and the simplicity of the model. Staging systems were compared using Harrell's concordance index. RESULTS: Median age was 70 years (range, 19-97years) and median follow-up time of 22 months (range, 1-250months). The median DOI was 6.0 mm (range, 0.1-70.0 mm). The five-year DSS rate was 80.7% (95%CI, 77.4-83.7%). We found significant association between DOI (hazard ratio, 1.21 [95%CI: 1.01-1.43]) and DSS on multivariable analysis. Based on a low Akaike information criterion score, improvement in the concordance index, and Kaplan-Meier curves, model 6 surpassed the AJCC staging system. CONCLUSIONS: Incorporation of DOI in the current AJCC staging system improves discrimination of T classifications in head and neck CSCC patients. LAY SUMMARY: The current staging system for head and neck cutaneous squamous cell carcinoma demonstrates wide prognostic variability and provides suboptimal risk stratification. Incorporation of depth of invasion in the T-classification system improves risk prediction and patient counseling. PRECIS: We propose improved head and neck cutaneous squamous cell carcinoma T staging that will include depth of invasion and should be considered in future versions of the American Joint Committee on Cancer after external validation.