Predictors of extensive subclinical spread in squamous cell carcinoma.

Abstract

e22093 Background: Extensive spread of squamous cell carcinoma (SCC) is often clinically imperceptible but has a profound impact on risk of recurrence and metastasis. Accurate identification of high-risk SCC enables educated choices in treatment modality and appropriate referrals for Mohs micrographic surgery (MMS). Methods: We conducted a retrospective cohort study of 1,019 patients with SCC removal by MMS at the Cleveland Clinic. We analyzed the relationship between various patient and tumor characteristics to aggressive subclinical extension (ASE), defined as tumors requiring >3 Mohs layers to obtain clear margins. Tumor locations were grouped into zones according to the American Academy of DermatologyÕs appropriate use criteria (AUC). Results: 67 of 1,149 tumors required >3 Mohs layers for removal. Tumor grade was the only statistically significant variable, with moderately-differentiated tumors having 1.828 times the odds of ASE compared to well-differentiated tumors (95% CI = [1.026, 3.255], p = 0.041). Poorly-differentiated tumors had 4.273 times the odds of ASE compared to well-differentiated tumors (95% CI = [1.755, 10.4], p = 0.001). Conclusions: Current treatment guidelines for SCC are largely derived from data on all non-melanoma skin cancers. SCC is associated with higher metastatic potential than basal cell carcinoma (BCC), emphasizing the importance of independent analysis of SCC. The power of this study was limited by the small number of extensively spread tumors. Despite this limitation, our analysis shows poorly and moderately-differentiated tumors are correlated with a higher rate of ASE compared to well-differentiated tumors. Moderately-differentiated tumor grade is not included in the AUC, however treatment of these tumors with MMS might improve patient outcomes and minimize costs associated with recurrence. Our results show risk stratification by zone has minimal utility in identifying ASE-SCC. Improved understanding of location-based differences in tumor behavior could reveal discrepancies between SCC and BCC, enabling revised zones specific to SCC. [Table: see text]