Study Population
The source population consisted of 49,270 individuals reported to FCDS and diagnosed between 2005–2018. Nearly 21% of total cases were excluded from analysis due to unknown smoking status at diagnosis (n = 10,280) leaving 38,990 cases in our final analysis. FCDS, the statewide cancer registry, collects data from multiple sources, including patient medical records and death certificates. These records are obtained from various healthcare facilities in Florida, such as hospitals, outpatient surgery centers, private physician offices, and radiation therapy facilities. To ensure accuracy, FCDS has undergone external audits, with a completeness level estimated to be over 95%. FCDS has met or exceeded the North American Association of Central Cancer Registries (NAACCR) standards of quality, timeliness, and completeness for every year since 1995.17
Inclusion Criteria
All individuals with site codes for lip, oral cavity, and pharyngeal cancer (C00.0-C14.8) as defined by the International Classification of Diseases for Oncology, Third Edition (ICD-O-3) were included in this study.18 The analysis included information on age, sex, race/ethnicity, socioeconomic status, insurance status, smoking status, stage at diagnosis, treatment, survival, and cause of death.
Study Variables
Primary cancer site was divided into fourteen different sublocation by ICD-O-3 code: lip (C00.0-C00.9), base of tongue (C01.9, C02.4, C02.8), anterior tongue (C02.0-C02.3, C02.9), floor of mouth (C04.0-C04.9), gum (C03.0, C03.1, C03.9), soft palate and uvula (C05.1, C05.2), hard palate (C05.0, C05.8, C05.9), cheek and other mouth (C06.0-C06.9), salivary gland (C07.9-C08.9), tonsil (C09.0-C09.9), oropharynx (C10.0-C10.9), nasopharynx (C11.0-C11.9), hypopharynx (C12.9-C13.9), and other oral cavity and pharynx (C14.0-C14.8) as in previous research.19 Socio-demographic variables comprised sex, age, race/ethnicity, SES, and insurance types. Clinical-pathological factors included smoking status at diagnosis, cancer stage, cancer location, and treatment received (chemotherapy, radiation, surgery). In patients with multiple forms of OPC, only the first diagnosed oral or pharyngeal tumor location was included in this study. Smoking status was categorized ever smokers, which included current and former, and never smokers, which was defined as having smoked less than 100 cigarettes in their lifetime. To qualify as a former smoker a patient must have quit smoking for over 1 year prior to diagnosis. Sex was determined based on information from medical records. Race/ethnicity was self-reported and categorized into the following mutually exclusive groups: non-Hispanic White (White), non-Hispanic Black (Black), Asian and Pacific Islander, American Indians, and Hispanics of any race. Cancer stage at diagnosis was defined using the Surveillance, Epidemiology, and End Results (SEER) staging categories, including localized, regional, distant, and unknown stages. Socioeconomic status was determined based on the percentage of residence living under the poverty level in the given census tract and was categorized as: very low poverty (0% to < 5%), low (5% to < 10%), intermediate (10% - <20%), high (20% to < 100%). Insurance type was classified as one of the following: private, Medicare, Medicaid, no insurance, or insurance status unknown.
Statistical Analysis
Socio-demographic and clinical characteristics were compared between never smokers and ever smokers using chi-squared test and Mood’s median test where appropriate. Cancer location by sex and smoking status was also analyzed. The study outcome was oral cavity and pharyngeal cancer-specific mortality. This was based on cause of death information obtained from death certificates and following specific SEER rules for cause specific cancer death for OPC as an only or one of multiple primaries.20 The observed survival time was calculated as the difference between the date of diagnosis and the date of death due to oral cavity and pharyngeal cancer or December 31, 2018, whichever occurred first. Patients who died of other causes or were alive at the end of the study period (December 31, 2018) were censored. Cox proportional hazards regression analysis was conducted to assess potential prognostic factors of survival in both univariable and multivariable models including sex, age, race/ethnicity, socioeconomic status, smoking status at diagnosis (never, current, and former), insurance type, cancer stage, primary cancer site, sequence factor number, and treatment (surgery, chemotherapy, and radiation).
To directly assess the prognostic role of cigarette smoking, we compared current smokers against former smokers as a reference in a model restricted to ever smokers, adjusting for potential confounders and covariates. This analysis was conducted because smoking is linked to significant underlying health conditions, such as chronic obstructive pulmonary disease, cerebrovascular disease, hypertension, and heart disease.21,22 These comorbidities can have a notable impact on the survival of individuals with OPC and can be present in both groups of ever smokers. By comparing current to former smokers, we can determine if there is an additional independent prognostic factor associated with current smoking that is not found in the former smoker population, beyond the smoking-related comorbidities that may be present in both groups.
The proportional hazard assumption was confirmed through graphical analysis and Schoenfeld residual tests. All tests were conducted as two-sided tests with a significance level (alpha) of 0.05. IBM SPSS version 28.0.0.0 was utilized for all statistical analyses.