Is Smoking Related to a More Aggressive Oral Cavity Squamous Cell Carcinoma with Poor Survival? A Multi-Institutional Collaborative Study

Background: There is recent overall trend towards an increase in the incidence of oral cavity SCC (OCSCC) most marked in oral tongue in young white non-smoker females. Tobacco smoking is a powerful risk factor for OCSCC; however, its influence on disease aggressiveness and survival is controversial. Objectives: To evaluate whether OCSCC in smokers have distinct clinicopathologic features and whether smoking status has any independent impact on survival outcomes. Methods: A collaborative database of patients with primary OCSCC among 6 academic medical centers was queried for all non-metastatic cases diagnosed between 2005 and 2015. All cases were treated with surgery ± adjuvant radiation therapy (RT) ± concomitant chemotherapy (CRT). Patients were categorized based on smoking history and the resultant groups were compared for pathologic features and treatment using t test and Chi-squared tests. Kaplan-Meier curves, Log-rank test and multivariate analysis (MVA) were used for recurrence free survival (RFS). Results: We identified 1055 cases that met our inclusion criteria after excluding 227 with unknown smoking history and inadequate follow up. The median follow up time was 40 months (15–195). Surgery alone, adjuvant RT and CRT were utilized in 32%, 37% and 31% respectively. Smokers either during or at any point before diagnosis (ever smoked) were 730 cases (69%), whereas never smoked constituted the remaining 31% (n=325). Ever smoked OCSCC was related to alcohol abuse (38% vs 3%) and was more located in the floor of mouth (28% vs 6%); whereas never smokers were predominantly females (61% vs 31%) with tumors more in oral tongue (30% vs 19%) (p\u3c0.001 for all). Rates for utilization of adjuvant RT and CRT were not associated with smoking status (p=0.14). Adverse pathologic features were similar in smoking groups including pT and pN stage, lymphovascular (LVSI) and extracapsular space invasion (ECE); except for more perineural invasion (PNI) (48% vs 37%; p=0.003) and less depth of invasion \u3e10 mm (73% vs 83%; p=0.01) in smokers. Smoking did not influence RFS with 2 and 5 years RFS (68% and 58% for ever smoked vs 67% and 56% for non-smokers; p=0.52). Never-smokers had worse RFS than smokers for AJCC stage I (p=0.03) and not for other stages. Even when we further categorized our study population into current, former and never smokers; and when we used smoking index, similar results were attained. On MVA, smoking was not associated with any effect and only higher stage, PNI, LVSI, ECE and positive final margins were independently associated with worse RFS (p\u3c0.05 for all). Conclusions: This collaborative database has one of the largest OCSCC cohorts treated using modern modalities with adequate F/U; smoking was not associated with any adverse risk features except for perineural invasion. Never smokers had more depth of invasion and worse survival in stage I which mandates further study including genetic analysis. Smoking failed to independently impact outcomes and only classic risk factors remained significant