2254 - Survival Rates Following Adjuvant Radiation in Cutaneous Squamous Cell Carcinoma with Perineural Invasion: An Analysis Using the National Cancer Database

Purpose/Objective(s): There remains a need to stratify high-risk cutaneous SCC (cSCC) patients with perineural invasion (PNI) and if they may benefit from adjuvant radiation (RT). This study evaluated overall survival (OS) rates of head and neck cSCC with PNI, with or without RT. OS rates of patients who also had Mohs micrographic surgery (MMS) and wide local excision (WLE) were also performed.

Materials/Methods: This retrospective cohort study used the National Cancer Database (NCDB) for adults with cSCC. Primary site codes C44.0-C44.4 and primary surgical site codes 34-36 (for MMS) and 45 (for WLE) were used. Patients with missing data on race/ethnicity, age, gender, or PNI were excluded. OS was analyzed using Kaplan-Meier product limit curves where if the endpoint event (death) did not occur, the number of months until the last follow-up where the patient was alive or lost to follow-up was used. The groups were compared using the log-rank test. The multivariate proportional hazards Cox regression model was used to determine if demographic variables were significantly associated with OS. Results were reported as hazard ratio (HR) with 95% confidence intervals. A subset study was also performed, looking at OS in patients who also received MMS or WLE with or without PNI, and with or without RT. p<0.05 was used for statistical significance. All analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC).

Results: There was a significant difference in race (White: 92% vs. 85.8% respectively, p=0.018) and the rate of RT (49.6% vs. 11.3% respectively, p<0.001) between those with PNI (n=276) vs. those without (n=653). There was not a significant difference in OS between subjects with and without PNI (log-rank test, p=0.1197). After adjusting for age and sex in the multivariate analysis, there was still no significant difference in OS between the two groups (HR =1.38; 95% CI: 0.98 - 1.96; p=0.0661). The interaction effect between PNI and RT was statistically significant in those with PNI, whereby the hazard ratio was less than 1 in both the univariate (HR=0.60; 95% CI: 0.34-1.06; p=0.0422) and multivariate analyses (HR=0.65; 95% CI: 0.37-1.15; p=0.0159). There was no statistically significant difference in OS when comparing patients with MMS (n=209) or WLE (n=142), with or without PNI or RT.

Conclusion: In this retrospective cohort NCDB study, the interaction effect in patients with PNI and who received RT when adjusted for age and sex indicated that death occurred at a slower rate. While there are several limitations to using a large retrospective cohort database, this data suggests a potential benefit to RT, with future studies needed to confirm our results.